AN AUTONOMOUS DIGGING VEHICLE  
 
 
Except where reference is made to the work of others, the work described in this thesis is 
my own or was done in collaboration with my advisory committee. This thesis does not 
include proprietary or classified information. 
 
 
______________________________________ 
Courtney A. Guasti 
 
 
 
 
 
Certificate of Approval: 
 
 
 
 
 
  
 
 
______________________________________ 
Ruel A Overfelt 
Professor 
Materials Engineering 
______________________________________ 
William F. Gale, Chair 
Professor 
Materials Engineering 
______________________________________ 
Jeffrey W. Fergus 
Associate Professor 
Materials Engineering 
______________________________________ 
Stephen L. McFarland 
Dean 
Graduate School 
AN AUTONOMOUS DIGGING VEHICLE  
 
Courtney Allen Guasti 
 
 
 
A Thesis 
Submitted to 
the Graduate Faculty of 
Auburn University 
in Partial Fulfillment of the 
Requirements for the 
Degree of 
Master of Science 
 
 
 
 
Auburn, Alabama 
August 7, 2006 
  iii 
AN AUTONOMOUS DIGGING VEHICLE 
 
 
Courtney Allen Guasti 
 
 
Permission is granted to Auburn University to make copies of this thesis at its discretion, upon 
the request of individuals or institutions and at their expense.  The author reserves all 
publication rights. 
 
 
 
 
____________________________ 
         Signature of Author 
 
 
 
 
____________________________ 
         Date of Graduation 
 
 
 
 
 
 
  iv  
VITA 
Courtney Guasti, son of Victoria Rhodes and John Aruzza, was born on March 7, 1979, 
in Burlington, Vermont.  He graduated Brooks High School in Killen, Alabama with an 
?Advanced Academic Diploma? in May of 1997.  He then attended Auburn University on 
scholarship where he obtained a Bachelors of Science in Textile Engineering in May 
2001.  After working for a year he then returned in August of 2002 to enter the Graduate 
School at Auburn University.  He married Abigail Gray, daughter of Don and Mona Gray 
of Killen, Alabama on July 16, 2005.  After finishing his coursework, Courtney was 
employed by the U.S. Army Aviation & Missile Research, Design and Engineering 
Center, located on the Redstone Arsenal in Huntsville, Alabama. 
 
 
  v  
THESIS ABSTRACT 
AN AUTONOMOUS DIGGING VEHICLE  
Courtney Allen Guasti 
Master of Science, August 7, 2006 
(B.S. Textile Engineering, Auburn University, 2001) 
127 Typed Pages 
Directed by Dr. William Gale 
 
This project is intended to develop a self-guided digging vehicle for various 
applications.  Hence, the main focus for this vehicle is to have the ability to dig through 
loose soils.  This project has focused on key digging technologies such as the use of 
shape memory alloys (SMAs) to motor power, precision motors to power conical and 
helical augers and biological movements as observed in ?fossorial? creatures. 
 Initially SMAs were selected as the source of motion to drive an auger to move 
the vehicle through the soil; whereas precision motors have become more promising for 
this purpose and became the second phase of this project.  It was observed that SMAs of 
a feasible geometry could not produce the torque required to move the vehicle.  Small 
SMAs have a low energy conversion efficiency (less than 5%) and cannot handle large 
loads.  Large SMAs can handle larger loads, however these seem to experience shape 
?amnesia? at an increased rate (with currents over 10 A), when compared with standard 
SMAs.  
  vi  
The use of an electric motor offers long lasting capabilities and when used with currents 
of around 15 A can produce the torque needed to bury itself sufficiently with customized 
augers. 
 The third phase of this project looks at researching digging creatures and their 
movements to move through the earth.  To scale down the amount of creatures involved 
in the research, an initial criterion of being ?fossorial? was established.  Upon selection of 
the creatures to be researched, specific questions were created to maximize the time 
researching each creature and not researching useless information.  A weighting system 
was established ranking the following criteria deemed important for a digging vehicle: 
speed, power, depth, material substance, mechanical complexity, geographic range and 
size of organism.  This ranking consisted of a scale of 1-5, with a score of one being 
harmful and a score of 5 being important.  Three scenarios were then created, giving each 
of the seven metrics different percentages based on the importance of each:  Scenario 1, 
small body (? 30 cm; 12 in) at low depths (? 152 cm; 60 in); scenario 2, small body (? 30 
cm; 12 in) at medium depths (152 cm; 60 in ? 610 cm; 240 in) and scenario 3, medium 
body (30 cm; 12 in ? 91 cm; 36 in) at deep depths (? 610 cm; 240 in).  Each creature was 
then given a final score and compared to each other, with the highest score from each 
scenario receiving the chance to be selected for further modeling.   
Although more information is needed for all the creatures to identify an animal 
most worthy of modeling, the available information indicates that the Townsend mole is 
the creature worthy of being selected. 
 
  vii  
ACKNOWLEDGMENTS 
 I would like to thank Dr. William Gale for his support and guidance throughout 
my graduate career.  The professional traits that Dr. Gale has taught me will be 
invaluable throughout my career.  Thank you Dave Lindahl for his patience and guidance 
throughout the many difficulties faced during my projects.  Thank you to Dr. Jeffery 
Fergus and Dr. Ruel Overfelt for their help, advice and guidance given.  I would also like 
to thank Ben Beyer, Robert Love and Chris Long for their relentless effort in helping 
during all aspects of the research. 
 I especially want to thank my wife for sticking with me during my graduate 
career, even though she thought I would never finish.  Without your help, I wouldn?t be 
where I am today.  I love you! 
 Last but not least, I would like to thank all my friends who have made my years at 
Auburn enjoyable and very memorable.  May we all meet up in the years to come.   
 
  viii  
Style manual or journal used:  Journal of Metallurgical Transactions A. 
Computer software used:  Microsoft Word 2000, Microsoft Excel 2000 
 
 
 
 ix
TABLE OF CONTENTS 
LIST OF TABLES ..........................................................................................................xiv 
LIST OF FIGURES ......................................................................................................... xv 
1. INTRODUCTION .....................................................................................................   1 
2. LITERATURE REVIEW ..........................................................................................   4 
 2.1 History of Drilling ............................................................................................   4 
      2.2 Earthmoving Equipment....................................................................................   5 
 2.3. Augers ...............................................................................................................   8 
 2.4. Revolutionary Drilling....................................................................................... 10 
 2.5. Shape Memory Alloys....................................................................................... 10 
  2.6. Mechanical Biomimetic Devices....................................................................... 13 
  2.7. Wildlife Research .............................................................................................. 16 
 2.8. Autonomous Technology .................................................................................. 20 
  2.8.1 Power Supply......................................................................................... 20 
  2.8.2 Single Vehicle vs Group Approach....................................................... 22 
  2.8.3 Communication ..................................................................................... 23 
3. RESEARCH OBJECTIVES ...................................................................................... 25 
4. MATERIALS AND METHODS............................................................................... 27 
 4.1 SMA Testing ..................................................................................................... 27 
 4.2 Auger Testing .................................................................................................... 33
  4.2.1 Electrical Motor..................................................................................... 30 
 x
  4.2.2 Conical Augers ...................................................................................... 30 
  4.2.3 Helical Augers....................................................................................... 33 
  4.2.4 Torque Requirements ............................................................................ 35 
  4.2.5 Push Test ............................................................................................... 36 
  4.2.6 Pull Test................................................................................................. 38 
  4.2.7 Colored Sand Experiment...................................................................... 39 
  4.2.8 Customized Auger ................................................................................. 41 
   4.2.8.1 Plain Sand Experiments............................................................. 42 
   4.2.8.2 Sand Mixture Experiments ........................................................ 43 
   4.2.8.3 Compacted Sand Mixture Experiments..................................... 44 
   4.2.8.4 Torque Requirements of Customized Auger............................. 45 
  4.2.9 Digging Energy Requirements .............................................................. 45 
 4.3 Vehicle Experiments ......................................................................................... 45 
  4.3.1 Earthworm Concept............................................................................... 45 
  4.3.2 Friction Reduction................................................................................. 46 
 4.4 Biomimetic Analysis ......................................................................................... 47 
  4.4.1 Brainstorming........................................................................................ 47 
  4.4.2 First Cut................................................................................................. 47 
  4.4.3 Initial Round of Questions..................................................................... 48 
  4.4.4 Second Round of Question.................................................................... 48 
  4.4.5 Weighting System ................................................................................. 49 
  4.4.6 Flow Chart for Biomimetic Design ....................................................... 50 
5. EXPERIMENTAL RESULTS................................................................................... 51 
 xi
 5.1 SMA Testing ..................................................................................................... 51 
  5.1.1 First Batch ............................................................................................. 51 
  5.1.2 Second Batch......................................................................................... 55 
 5.2 Auger Testing .................................................................................................... 58 
  5.2.1 Torque Requirements ............................................................................ 58 
  5.2.2 Push Test ............................................................................................... 59 
  5.2.3 Pull Test................................................................................................. 60 
  5.2.4 Colored Sand Experiments .................................................................... 61 
 5.3 Customized Auger ............................................................................................. 63 
  5.3.1 Plain Sand Experiments......................................................................... 63 
  5.3.2 Sand Mixture Experiments.................................................................... 64 
  5.3.3 Compacted Sand Mixture Experiments................................................. 65 
  5.3.4 Torque Requirements for Customized Auger........................................ 66 
 5.4 Vehicle Experiments ......................................................................................... 66 
  5.4.1 Earthworm Concept............................................................................... 66 
  5.4.2 Friction Reduction................................................................................. 67 
 5.5 Biomimetic Analysis ......................................................................................... 69 
  5.5.1 Weighted Results................................................................................... 69 
6. DISCUSSION............................................................................................................ 71 
 6.1 Shape Memory Alloys....................................................................................... 71 
 6.2 Auger Testing .................................................................................................... 72 
  6.2.1 Conical Augers ...................................................................................... 72 
  6.2.2 Helical Augers....................................................................................... 73 
 xii
  6.2.3 Custom Auger........................................................................................ 74 
 6.3 Vehicle Experimentation................................................................................... 75 
  6.3.1 Earthworm Concept............................................................................... 75 
  6.3.2 Air Hose/Friction Reduction ................................................................. 75 
 6.4 Biomimetic System ........................................................................................... 76 
7. CONCLUSIONS........................................................................................................ 78 
 7.1 Shape Memory Alloys....................................................................................... 78 
 7.2 Augers ............................................................................................................... 79 
  7.2.1 Conical Augers ...................................................................................... 79 
  7.2.2 Helical Augers....................................................................................... 79 
  7.2.3 Custom Auger........................................................................................ 80 
 7.3 Biomimetic System ........................................................................................... 81 
8. RECOMMENDATIONS FOR FUTURE WORK .................................................... 83 
REFERENCES ............................................................................................................... 85 
APPENDICES ............................................................................................................... 90 
Appendix ? A: Draft CAD Drawing of Conical Auger ................................................. 91 
Appendix ? B: Auger and Motor with Fin Design......................................................... 93 
Appendix ? C: Extended Auger..................................................................................... 95 
Appendix ? D: Draft CAD of Custom Auger ................................................................ 96 
Appendix ? E: List of Initial Question........................................................................... 97 
Appendix ? F: Scenarios for Weighted Values ............................................................. 99 
Appendix ? G: Biomimetic Flowchart...........................................................................100 
Appendix ? H: SMA Spring Data ..................................................................................101 
 xiii
Appendix ? I: Final Results for all Creatures ...............................................................103 
Appendix ? J: Information Found for Each Creature ...................................................104 
Appendix ? K: Metric Results for the Townsand Mole.................................................106 
 
 
 xiv
LIST OF TABLES 
 
Table 1 SMA Conversion Efficiency of the First Batch of Springs ..................... 51 
Table 2a Initial Observations of a SMA Spring at Low Currents and Loads......... 56 
Table 2b Second Testing of a SMA Spring at Low Currents and Loads................ 56 
Table 3a  Digging Results of Plain Sand Experiments............................................ 63 
Table 3b  Estimated Digging Energy of Plain Sand Experiments ........................... 64 
Table 4a  Digging Results of Sand Mixture Experiments ....................................... 65 
Table 4b  Estimated Digging Energy of Sand Mixture Experiments ...................... 65 
Table 5a  Digging Results of Compacted Sand Mixture Experiments.................... 66 
Table 5b  Estimated Digging Energy of Compacted Sand Mixture Experiments ... 66 
Table 6 Torque Requirements of the Custom Auger ............................................ 67 
Table F-1 Weighted Values for Three Scenarios ..................................................... 99 
Table H-1  Results of Spring 1 with 44.5N at Various Currents ...............................101 
Table H-2  Results of Spring 2 with 44.5 N at Various Currents ..............................102 
Table H-3  Results of Spring 3 with 26.7 N at Various Currents ..............................102 
Table I-1  Final Results for All Creatures ................................................................103 
Table J-1 Information Found for Each Creature......................................................104 
Table K-1  Metric Results for the Townsand Mole ...................................................106 
 
 xv
LIST OF FIGURES 
 
Figure 1  - Hysteresis Loop in Shape Memory Alloys [18] ..................................... 13 
Figure 2  - Testing Platform ..................................................................................... 28 
Figure 3  - Machined Attachment Connecting the Conical Auger to the Motor...... 31 
Figure 4  - Motor with Attached Rod for Counter Rotation..................................... 32 
Figure 5  - Coupling Devices Connecting the Helical Auger to the Motor.............. 33 
Figure 6  - Adaptor Connecting the Helical Auger to the Motor ............................. 34 
Figure 7  - View of Augers Tested ........................................................................... 35 
Figure 8  - View of Aluminum Rods used for Push Test ......................................... 37 
Figure 9  - View of Pull Test Pulley System............................................................ 39 
Figure 10 - Multi-Colored Sand used to Examine the Auger Sand Interaction ........ 40 
Figure 11 - View of Auger Designed by UMR ......................................................... 42 
Figure 12 - Diagram View of Earthworm Design..................................................... 46 
Figure 13 - Graphical Representation of the Current vs. Stroke Results .................. 54 
Figure 14 - Graphical Representation of the Force vs. Time Results ....................... 55 
Figure 15 - Graphical Representation of the Depth vs. Torque for Various Augers. 59 
Figure 16 - Graphical Representation of the Results from the Push Test ................. 60 
Figure 17 - Profile of Digging Trial using a Sand Core............................................ 62 
Figure 18 - Stopping Point of the Earthworm Digging Trials................................... 68 
Figure 19 - View of Motor with Air Hoses ............................................................... 69
 xvi
Figure A-1 Side and Top View of Conical Auger................................................... 91 
Figure A-2 Cutaway View of Conical Auger.......................................................... 92 
Figure B-1 Motor with Three Small Fins................................................................ 93 
Figure B-2 Motor and Conical Auger with Two Large Fins................................... 93 
Figure B-3 Motor and Conical Auger with Rod Fin Design................................... 94 
Figure C-1 Extended Auger used for Pull Test ....................................................... 95 
Figure D-1 Side View of Custom Auger................................................................. 96 
Figure G-1 Biomimetic Flowchart ..........................................................................100 
 
  1  
1. INTRODUCTION 
Screenwriters and storytellers conjure up fairytales about self thinking and self 
assembly robots that wreak havoc and destruction upon mankind in order for machines to 
gain their freedom and eventually rule the world.  Although this scenario is still far from 
becoming reality, science and technology advances every generation, from the invention 
of automobiles to ease travel on the ground to the birth of aviation, which enables us to 
soar through the air.  When ground and sky were no longer enough, space travel opened 
up a much wider and more complex playground for us to explore.  However, while all 
these things have helped mankind become faster, stronger and better, a new wave of 
technology has grown to be part of our lives: remote control.  Mankind can now reap the 
benefits of older inventions, but without having to be physically present atop the 
machine.  Today, those same machines that were originally remotely controlled are now 
being upgraded to give them the capacity for what might very loosely be described as 
?intelligence?.  Autonomous vehicles are being created that are capable of covering 
ground, air and sea. 
Large earthmoving machinery has followed the same progression.  Today?s 
machines have sensors and computer controlling devices instead of human operators that 
enable them to react with the surrounding environment [1].  Large amounts of earth can 
be moved with these machines, but not without overcoming large forces.  Instead of 
moving the earth out of our way, why not move through the earth?  
  2  
Augers are widely used in the construction industry to accomplish simple tasks, such as 
post hole digging, or even large tasks, such as drilling roadways through mountains or 
under rivers.  An auger?s design is intended to break up the material and transport it out 
of the way, hence moving within the earth and moving the earth, simultaneously.  
Unfortunately, these uses require power supplies that are more substantial than a few AA 
batteries and forces larger than a few Newtons. 
A closer look at Mother Nature reveals that insects and some mammals have no 
problem moving within the earth.  They need no large power supplies or (at least in some 
cases) hours of work to travel distances of five or even ten times the length of their 
bodies.  What is it that gives these creatures the ability to dig so effectively at depths that 
are many, many times their own size?   
Just as autonomous vehicles in the sky, sea or on land have their own unique 
functions and capabilities, an autonomous digging vehicle has the potential to be used in 
many real world applications, such as civil search and rescue missions, geological 
exploration on earth and other planets, and industrial uses such as drilling air vents for 
mine shafts or wire and cable emplacement for construction projects or data 
transmissions.   
This thesis consists of a study of three different approaches to designing an 
autonomous digging vehicle, namely the possibility of creating one through the use of 
shape memory alloys (SMA), augers, or new techniques inspired by Mother Nature?s 
own creations.  The project originally intended to have the vehicle move through a given 
medium, in this case sand, through the use of shape memory alloys.  This metamorphosed 
into the possibility of a vehicle constructed of a motor and auger.  The final phase 
  3  
involved a literature study and observant evaluation of several creatures and their 
different digging techniques in order to examine how their physical characteristics 
contributed to their outstanding digging performance. 
  4  
2. LITERATURE REVIEW 
2.1. History of Drilling 
Since the earliest days of pre-history, mankind has drilled holes in the earth?s 
crust.  Drilling equipment and needs, from today?s standpoint, can be grouped into 
different categories based on the specific needs of different industries, such as those 
associated with gas, oil or mineral exploration, construction and excavation or quarrying 
and mining [2].  However, in ancient times, the main need was to find a source for water 
and some of the best-preserved structures from ancient civilizations are water wells.  One 
example, dug in ancient times, is located in Cairo, Egypt and is named ?Joseph?s Well? 
[3].  With a total depth of about 90 meters (295 ft) and traditionally powered by oxen 
pulling on long chains with buckets attached, records show it to have produced water 
right up until the year 1900 [3].  An early European well, located outside the village of 
Lilliers in northern France, was drilled in 1126 A.D. and still produces water to the 
present day [3].  One of the oldest forms of drilling dated back around 600 B.C. when the 
Chinese developed a technique that is now called percussion drilling [4].  Chung, [3] 
describes the early process as a spring pole, tapered on one end, being placed on a 
fulcrum, giving the effect of a lever.  A heavy weight is placed on the thick end and the 
drilling tool is attached by rope to the tapered end.  A hollowed piece of wood is placed 
in the ground, directly beneath the cutting tool to provide stability to the borehole and 
guidance for the cutting tool.   
  5  
A group of two men then lifts and lowers the cutting tool inside the log, with the help of 
the fulcrum, giving the tool a spin with each strike to make sure the bore hole is even and 
symmetrical. This work continues day and night, for many years if necessary, until the 
final depth is reached.  Another form of drilling, rotary drilling, was used as far back as 
3000 B.C. by the Egyptians [5].  Here, a sharp edged flint is attached to one end of a 
forked stick.  The tool is then spun into the bore-well with the leverage of the forked end.  
Helical grooves can be seen along the walls of the boreholes produced using this method 
[3].  Both percussion and rotary drilling are still used in modern times, although the 
materials have advanced for percussion drilling, with the spring pole and rope being 
replaced by steel pipe and chains and the striking against the ground has become 
automated with the use of gears and machines.  A French engineer named Leschot [6] 
developed the first modern rotary drill around the 1960s.  Constructing a tunnel for a 
railway system linking Modane, in France, and Bardonecchia, in Italy, Leschot built a 
belt driven rotary drill with a black diamond face [3, 6].  Percussion drilling was used for 
oil drilling in the early days of the oil industry, although it was eventually replaced by 
rotary drilling in the 1930s due to the faster rotations that were possible.  Before the 
1930s, speeds of only 30-60 rpm could be maintained but with the introduction of 
diamond bit sets, speeds increased dramatically to 200-400 rpm [3].   
2.2. Earthmoving Equipment 
One of the leading industries that makes extensive use of earthmoving equipment 
is construction.  There will always be a demand for bigger and better skyscrapers and 
bridges, and hence a need for bigger and better earthmoving vehicles.  It is not the 
vehicles that move the earth but the cutting tools attached, which in the case of bulldozers 
  6  
and power shovels takes the form of a scoop or bucket.  Experiments designed to optimze 
excavator buckets have tested everything from the width of the bucket to the number of 
teeth and the spacing between each, the length and position of the teeth or buckets with 
no teeth at all [7].  Maciejewski et al. [7] concluded that the total amount of work done 
increased as the number of teeth increased and stabilized upon reaching a high number of 
teeth.  These types of fundamental experiments are being performed in order to provide a 
better understanding of a soil?s behavior.  Experiments have shown that the initial 
conditions (blade geometry, soil type) and the operating conditions (cutting speed and 
depth) have a significant effect on a machine?s productivity [8].  
Doing field experiments can be tedious and expensive over a period of many 
trials, but the use of Finite Element Analysis (FEA), can help ease this burden.  FEA 
simulations have the advantages of not depending on the accuracy of measuring devices, 
as in physical experimentations, replicating more sophisticated material phenomena and 
revealing details of the soil-blade interactions [8,9].  Simulations can be set up with 
different variables, such as cutting speeds and depths that are not limited by the available 
machinery and space. However, the limiting factor for FEA models is the computational 
power, as an increase in mesh density causes a considerable increase in computing time 
and power [10].  Many models are being investigated, although there is some difficulty in 
providing realistic estimates of soil strength.  Some research, like that of Georgiadis et al. 
[9], are helping to validate the empirical models for soil shear strength by comparing the 
failure models with experimental results.   
Another focus on earthmoving equipment has been to optimize the vehicles 
themselves.  The approach generally taken has been to provide a continuous monitoring 
  7  
system to assist the operator.  However Maciejewski and Jarzebowski [1] are attempting 
to develop an automated vehicle whereby the vehicle can react to soil conditions and 
optimize the earth-working process.  Again, by experimenting with the shape and 
trajectory of the excavating bucket they concluded that the digging method of cutting thin 
layers of soil, as performed by bulldozers, is inefficient as this method involves a higher 
energy dissipation per unit volume when compared with other digging methods.   
To help reduce the need for bulky earth-moving equipment, the use of trenchless 
technology is greatly increasing.  The gas line industry has invested considerable research 
time to developing better vehicle designs specifically for directional drilling and 
trenchless technology [11].  With increasing worry over environmental issues such as 
animal habitat impact and disruption, directional boring vehicles offer many advantages 
compared to conventional earthmoving vehicles (bulldozers, cranes, scrapers) [12].   
These machines are more cost-effective as the space and time to set up is minimal with a 
corresponding increase in work productivity, maneuverability and number of bores per 
day [11].  
Research has also increased the capabilities of tunnel boring machines or TBMs, 
which are primarily used to drill tunnels in hard rock [13].  These machines are now 
successfully drilling in ground and rock conditions that were thought to be impossible 
years ago.  The first TMB built in the United States was in 1856, although ?neither 
metallurgy nor structural design were to the point where it could be successful in hard 
rock? [13].  However, technology has come a long way and today TBMs are very cost 
effective due to their ability to apply a relatively constant power, gather up the cuttings 
and convey them to a haulage unit.  With today?s emphasis on safety when tunneling, 
  8  
TBMs offer improved overhead protection during operations, creating smooth walls for a 
more stable opening and fewer collapses [13].  In terms of automation, TBM companies 
are experimenting with laser guidance to help in automatic steering [13,14].  By 
incorporating sensor systems, TBMs are also beginning to gain an ability to respond to 
changes in rock and ground conditions in order to establish an optimum machine 
performance and thus allow for maximum penetration [14].  As with all large vehicles, a 
smaller version may often be needed for different scenarios, hence the creation of small 
boring units or SBUs.  SBUs were developed in part because sometimes the size or cost 
of TBMs prevents their use.  SBUs can be used for smaller jobs such as ventilation 
tunnels or man-ways in mining [14,15].  Being smaller, they can also be steered more 
readily.  One such development by a Japanese contractor has created a TBM that can 
make a 90? turn.  This particular model creates its own access shaft, which not only 
shaves months off the time needed, but may also cut the project costs by as much as 15% 
[15]. 
2.3. Augers 
For the geological surveys required for constructing roads and large building 
complexes, soil sampling is a large-scale operation and with the new demands of space 
exploration, extra-terrestrial soil will be the next step.  Soil sampling consists of 
collecting samples from below the surface to help identify changes in their 
characteristics.  Augers have been used for this purpose for decades, although their 
earliest uses were for wood working.  Augers come in many different sizes, ranging from 
small hand held tools to large augers that are anchored on the back of trucks.  Augers are 
designed with spiral flights to keep material moving to the surface and hence do not 
  9  
require a continuous flushing medium to keep debris clear of the cutting tip.  Research by 
Sellmann and Brockett [2] has shown that a well designed auger bit can operate 
effectively in many different materials, as long as the necessary torque and thrust is 
applied, although some augers are designed for specific applications.  However, some 
disadvantages when using augers for drilling in harder materials are rapid bit wear and 
the large drill rig needed to provide enough thrust for holes nine inches or larger in 
diameter. 
As the National Aeronautics and Space Administration (NASA) pushes forward 
to explore our solar system and its planets, drilling and excavation technology will be a 
vital tool.  ?To date, drilling has only been conducted on the moon to the depth of a few 
meters and on Mars to depths measured in centimeters? [16].  Research and testing is 
underway, sponsored by NASA, for a tethered hole drilling system to help overcome 
these limited depths.  The grinding motion or shearing action of a cutting tool 
accomplishes the majority of rock cutting or down-hole drilling processes and a large 
thrust and/or torque is usually required, also known as ?weight on bit,? to keep the 
cutting tool continuously moving [16].  This is accomplished by anchoring the drilling 
mechanism above ground or, in the case of oil drilling, the weight on bit is self-provided 
by the weight of the drilling unit, which incorporates long tool strings.  Hill?s group [16] 
has developed, a Low Reaction Force Drill (LRFD) to address NASA?s needs, which 
consists of a ?detached, self-driven underground autonomous tethered drill system?.  
Some of the advantages of the LRFD are its low energy requirements, unlimited depth 
capability, and the ability to self-advance, which are all self-contained [16].  This requires 
less equipment and a correspondingly lower mass to transport, a major asset for NASA 
  10  
when considering the enormous cost of lifting equipment out of earth?s gravity well.  The 
LRFD uses a cutting bit at the tip of the device with multiple augers or helical flights 
acting as cutters to advance the drilling system [16].   
2.4. Revolutionary Drilling 
Along with augers, diamond bits and other traditional drilling devices, researchers 
are designing and testing new methods to dig deeper and faster than ever before.  One 
such method is not new, but has simply become more feasible with the advancement of 
technology, namely lasers.  In laboratory tests, lasers were aimed at pieces of sandstone 
and were able to cut through a 150 mm specimen in four seconds, ten times faster than 
rotary drills [17].  However, as rapid as lasers are, they also require a large amount of 
energy and would not be very efficient in large hole operations.  Another method being 
tested by the U.S. Department of Energy is thermal spallation [17].  Thermal spallation 
uses a heat source to quickly heat a rock so fast that it expands and explodes into small 
pieces.  Heat sources being experimented with include electric arcs, microwaves and 
simple flames.  The disadvantage is that spallation is limited to rocks with large crystals 
that fracture cleanly [17].  One of the more interesting methods is a mechanical device 
that melts its own way down.  Named the subterrene, this device is a large electrical 
resistor with a cone-shaped ceramic tip.  The disadvantage is its limited speed of about 
one meter per hour [17]. 
2.5. Shape Memory Alloys 
Shape Memory Alloys (SMA) are alloys that experience a Shape Memory Effect 
(SME) due to phase transformations.  Discovered in 1932 by Swedish physicist Arne 
Olander [18], the SME behavior was observed in a gold and cadmium alloy that could be 
  11  
plastically deformed at cool temperatures and then be promptly returned to the original 
shape through the application of heat. 
Shape memory alloys are constantly being sought, with about a dozen already 
proven.  Materials experience the SME due to a ?temperature and stress dependent shift 
in the material?s crystalline structure between two different phases called martensite and 
austenite? [18].  Most SMAs are thermoelastic in that it can be plastically deformed when 
at temperatures below the transition temperature and recovered when heated to 
temperatures above the transition temperature.  A shape memory alloy?s capabilities are 
due to a solid-state phase change in the material; ?a martensitic ? austenic phase 
transition - and at low temperatures ? to martensitic twinning? [19].  At lower 
temperatures the material is found in a twinned martensitic state.  With an applied stress 
the material can be deformed in which the martensite twin variants shear into a parallel 
orientation [19].  After removal of the stress, the structure remains until the material is 
heated, in which the sheared twins revert back into the organized austenitic phase.  Upon 
cooling the martensitic twins form again, completing the SME cycle.  Austenite?s cubic 
structure and un-deformed martensite?s twinned structure are the same size and shape on 
a macroscopic scale resulting in no change visibly [19, 20].  In contrast to this process is 
an SMA that is designed to work using superelastic behavior, that is, a stress induced 
martensitic formation [20].  Through changes in alloy composition and heat treatment, an 
SMA?s phase transformation temperature can be manipulated [20,21].  By controlling 
this temperature to be slightly below room temperature, the SMA will be in its parent 
austenitic phase even when no load is present.  When a load is applied however, a 
martensitic transformation will occur.  Because martensite has formed above its transition 
  12  
temperature, if the load is removed, the new phase becomes unstable and reverts back 
into the austenite phase and again, the original un-deformed shape is recovered.  This is 
different from SMAs that have a transition temperature above room temperature because 
no temperature change is necessary for the material to recover its original shape 
[19,20,21].  SMAs are typically heated up using one of two ways, either electrical current 
(resistance heating) or with a high temperature medium surrounding the material such as 
air or water [22].   
SMAs were first used as fasteners that functioned as coupling devices for piping 
systems and electrical connections [20].  Later, research focused on using the alloys as 
actuators for temperature-regulated systems such as greenhouse windows, automobile fan 
clutches and air conditioners [23].  Recent research has examined their use as actuators 
for micro-robotics.  Conventional actuators, such as DC, pneumatic and hydraulic 
motors, are not energy efficient due to their large volume and heavy mass [18].  SMAs 
have an advantage over these through their high force to weight ratio [18], high power 
density, large stress and strain [23] and noiseless operation capabilities [18], although, the 
actual electrical energy to mechanical work conversion can be very inefficient with 
SMAs. 
The main disadvantage for SMA applications is the cooling required.  For large 
actuation forces, larger SMA wires are needed, requiring larger power demands to heat 
and, in turn, a longer cooling time.  The dominating factor in a cycle is the cooling time.  
An SMA cycle follows a hysteresis loop (figure 1).  Starting at point 1, the material is 
completely martensitic at temperature M
F
.  Following the bottom curve, the temperature 
increases and austenite starts to form at temperature A
S
.  Further heating transforms the 
  13  
microstructure to complete austenite (A
F
), at point 2.  Upon cooling, the material follows 
the upper curve, in which martensite starts to form upon reaching temperature M
S
 and 
completes the loop, cooling back down to the 100% martensite microstructure at point 1.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 1.  Hysteresis loop in shape memory alloys, after Mavroids et al. [18]. 
 
2.6. Mechanical Biomimetic Devices 
Biomimetics is not a new field of science, but simply a new name given to help 
identify a particular engineering concept.  In fact, science has tried to replicate nature for 
hundreds of years.  Simply, biomimetics is the ability to apply biological concepts to 
engineering processes or systems.   
1 
Temperature 
% Austenite
 
% Martensite 
cooling 
heating 
M
S
 
M
F
 
A
S
 
A
F
 
2 
0 
100 
100 
0 
  
  14  
In the mid 1400s, Leonardo da Vinci drew sketches of flying contraptions based 
on the flight habits of birds [24].  However, it was not until the 1800s that the 
visualization of da Vinci?s drawings became reality with the construction of omnithopters 
[25], which  were flying devices.  Again based on the flight of birds, these mechanical 
devices had large feathered wing-like structures on both sides, powered to flap under the 
control of the human pilot.  Unfortunately, these devices quickly proved that there is 
more to flight than just wings flapping; birds are at ease with the motion due to their 
size/weight to wing span ratio.  When undertaken on a scale sufficient to lift a man and 
machine however, the concept proved to be unworkable [25].   
During the 1950s, an avid hiker by the name of George de Mestral noticed small 
?burrs? that clung to his pants and his accompanying dog?s coat.  A closer inspection 
under a microscope revealed that the burrs had small, stiff, hooks with which they could 
grab onto the looped fibers of his clothes and his dog?s hair.  From this observation came 
his invention of Velcro, whose name originated from combining the words velour and 
crochet [26].   
In the 1990s, high performance swimsuits were introduced that mimicked the skin 
of sharks.  This concept was based on observations of a reduction in the number of eddies 
in the water flow along the surface of the shark?s skin as it moved through the water.  The 
result was to copy the v-notch pattern and apply it to full body swimsuits.  This design is 
believed to reduce the drag encountered while swimming by up to 8% [27].   
One of the most well known biomimetic devices is the MIT RoboTuna, built by then 
Ph.D. candidate David Barrett in 1995.  This research focused on developing and 
studying a better propulsion method for underwater vehicles [28].  By mimicking the 
  15  
fastest fish ? the blue-fin tuna ? the group at MIT hoped to control the vortices 
encountered while moving underwater.  Normal vehicle movement, due to a screw or 
propeller is slowed by these vortices, and generally designers aim to eliminate them.  
However, fish intentionally cause and control with their fins these vortices to boost their 
speed.  Besides increased speed, the RoboTuna can operate at greater depths than most 
Autonomous Underwater Vehicles (AUV) because the skeletal structure does not require 
a pressurized chamber.  This will not only be beneficial in the study of thermal vents 
along the ocean floor, but also for other dangerous explorations or even search and rescue 
missions that maybe undertaken in the near future. 
Autonomous vehicles are currently a major focus for science and industry, with 
the Department of Defense leading the way in spending.  The Robotics Institute at 
Carnegie Mellon University has seen its federal funding more than double since 1994.  
Other well known technology based universities have also seen a 50% increase in just the 
last few years [29].  Some see a robotic nation in the near future and with the way the 
Defense Department is spending, the idea is not so far fetched.  The Pentagon is thought 
to have spent $3 billion on Unmanned Aerial Vehicles (UAV) between 1991 and 1999, 
with an estimated budget of $10 billion through the next 5-6 years [30].  In 1946 it was 
revolutionary to fly two, unmanned B-17 bombers from Hawaii to California by remote 
control from the ground, yet in 1999 when a completely computer controlled UAV flew 
itself across the Pacific to Australia where it landed perfectly [31], this feat was not 
considered revolutionary but rather expected, due to the natural progression of 
technology.  UAVs are not only confined to airplanes; unmanned helicopters are also 
being developed.  A group at MIT has developed a small, agile and acrobatic helicopter 
  16  
that they expect to be widely utilized.  The main driving force for this rapid development 
is for military applications, namely reconnaissance, weapons delivery in mountainous, 
urban or other hazardous regions, low altitude or tight spaced missions, all of which can 
send images to a remote location or to an aircraft already in flight.  Along with military 
applications, the MIT design group expects the helicopter to be used in surveying disaster 
sites or even filming aerial views for motion pictures [32]. 
The Defense Department is pushing for one-third of its ground vehicles to be 
unmanned by 2015 [30].  With this comes a challenge to industry and academia to 
produce the technology needed to control Autonomous Ground Vehicles (AGV).  The 
Defense Advanced Research Projects Agency (DARPA) has issued a challenge with a 
grand prize of $2 million to the team that can produce an unmanned ground vehicle 
capable of tackling different terrain and obstacles on its own.  Held annually, DARPA 
hopes to ?accelerate research and development in autonomous ground vehicles that 
someday will help save lives on the battlefield? [33].  In 2005, only 25 of the 86 groups 
that applied were selected to participate in the event, and only 15 of theses were allowed 
to proceed after initial qualification, inspection and demonstration stages.  However, so 
far no vehicle has successfully navigated the complete course and the prize remains 
unclaimed. 
2.7. Wildlife Research 
Overall, about 40 different creatures have been investigated by this project team, 
in terms of their digging capabilities.  Many share the same digging approach, however 
some distinguish themselves due to their absolute and relative tunnel lengths. 
  17  
Marine creatures are not the first subjects to come to mind when looking for a 
natural digger to model an autonomous vehicle after.  Some creatures, such as the razor 
clam, have an advantage compared to those on land because of the nature of their digging 
medium.  The razor clam forces its foot-like muscle into a soft substrate, in this case 
sand, and uses the ?foot? as an anchor by filling it with blood and expanding it.  Once 
secured in the sand, the muscles pull the shell down to the foot.  The two halves of the 
shell expand, compacting the surrounding material and the steps are repeated until the 
clam is safe from harm.  The razor clam can bury itself quickly, in as little as 15 minutes 
[34], compared to the geoduck clam, which digs in the same manner but can only dig a 
total of about 910 mm (36 in) in a period of three years [35].  Boring Clams take a harder 
path to security as they bore into something more stable, such as living coral.  After the 
clam secures itself to a hard substrate, an acid is released to break down the carbonate in 
the coral and the two shells start to grind away, slowly making an entrance [36].  The so-
called shipworm, which is actually a mollusk, is a relative of the boring clam family in 
that it also has small rigid shells that is uses to grind away material; however, here wood 
is the primary source instead of rock.  The shipworm gains nutritional support from the 
wood substrates as well as protection.  The tunnels remain open due to a calcium 
secretion along the tunnel walls [37]. 
Insects have also become a resource for those studying burrowing mechanisms.  
The problem is to find a true burrowing insect, as opposed to one that digs as a nymph for 
protection to carry out the remaining of its life cycle.  One true burrowing insect family is 
the mole cricket.  Because of the damage caused by the mole cricket, a group at the 
University of Florida has devoted a lot of time to researching and observing the mole 
  18  
cricket specifically.  These small insects, related to crickets and grasshoppers [38], have 
large paws as nymphs and adults that are essentially replicas of those found in mammals.  
The family can burrow with ease through sandy soils by parting the soil with their large 
claws, much as a mole would, and feed on the plants and grassland.  An entire database is 
devoted to these creatures [39].  The dung beetle is another insect that is a pure digger.  
The dung beetle is actually a family of beetles, as there are thousands of species.  These 
are an example of an environmentally friendly insect, feeding off the dung of animals.  
By extracting the liquid from the dung, the insects satisfy all of their nutritional needs 
including water [40].  The beetles can then roll up the remaining material and bury this to 
provide a habitat for their young.  Most species of dung beetle, however, dig tunnels 
underneath the dung pile and bring sections of it down into the tunnels [41].   Many 
insects, such as the June bug or cicada, will bury themselves as grubs (a younger stage in 
the life cycle) and stay buried for months until they mature into an adult.  The motion is 
that of a wiggle as their bodies squirm back and forth in the material until a satisfactory 
depth is reached, usually sufficient to keep them warm during the winter.   The insects 
mentioned so far are of a solitary nature, however the termite is well known for the large 
numbers found in each colony.  It is not known if termites work together in a coordinated 
fashion or if they randomly dig together. Termites are renowned for their wood eating 
abilities and live in large colonies underground, up to 5.5-6.0 m (18-20 ft) below the 
surface [42], or in large dirt mounds found above the ground.  The workers have large 
heads with massive pincers that they use to tear off pieces of wood or carry grains of soil.   
An interesting creature that digs in its own unique manner is the trapdoor spider.  This 
particular spider creates burrows for protection, a place to mate and a place for their 
  19  
offspring.  They loosen soil with small, hard, spike-like hairs attached to muscles called 
rastellum [43].  These spiders are able to maneuver the rastellum back and forth against 
the soil, breaking up small amounts at a time.  They then remove the soil in one of two 
ways, either by pushing the loose soil out with their legs or by forming clumps of soil that 
are wrapped in silk and carried out as a package.  The burrows protect the spiders from 
temperature fluctuations, flooding and caving in due to a silk lining around the walls.   
One amphibious creature that burrows is the Mexican burrowing toad, which 
prefers areas with loose soil [44] and digs by using the spades found on their powerful 
legs [45].  However, there is no information available on the exact motion used to burrow 
with their spades.   
The bulk of the research in this area has come from mammals, specifically 
families of rodents, such as squirrels, gophers and gerbils, moles and mole rats.  Rodents 
dig in the usual manner, with their legs in the front of their body.  Some, like the plains 
pocket mouse and the yellow cheeked vole, live in shallow burrows of 0.30 m (12 in) 
[46] or less while others, for example the olive backed  pocket mouse and Malagasy 
Giant rat live in burrows that reach down 2 m (79 in) [47]or deeper.  The only difference 
between the many burrowing squirrels and gerbils is the depths of their burrows. 
Gophers are highly fossorial and have bodies built for their living style, with 
?thick, tapering bodies, short limbs with strong claws, loose skin, and reduced eyes and 
ears? [48].  They also have large and ever-growing incisors, which they use to scrape at 
the soil to loosen it up.  After a considerable amount of soil is accumulated, they hold it 
with their forearms against their chest and push it out to the surface with their bodies 
[49].   
  20  
Mole rats are a social group living in colonies that coordinate their effort when 
creating burrows.  They have tiny eyes that can detect light and nothing else, a wide and 
flat head and, like gophers, large incisors that are used to scrap at the soil.  The lead mole 
rat will loosen the soil with its teeth and then push the loose material underneath and 
behind its body with its feet.  A train of mole rats behind the lead digger will then 
continuously push the material up and out of the tunnel to the surface by passing it to the 
next mole rat behind them.  In the naked mole rat, 25 % of their muscle mass is located in 
their jaw [50], enabling constant digging with their teeth.   
The most impressive digging animal is the mole, as both its claws and body are  
highly adapted for burrowing.  The forelimbs of the mole are located to the sides of its 
body, instead of in front, and are pointed outward, as if ready to swim through the soil.  
The mole?s humerus bone and muscle structure is uniquely adapted to allow the 
forelimbs to rotate in this manner while combating the forces encountered as they push 
the soil away from the body.  Their bodies are almost cylindrical, as well as streamlined.  
Like the gopher, the eyes and ears are small, forming a robust and solid skull.  The claws 
on the forelimb are as wide as they are long, forming a ?broad and shovel-like? [51] 
structure that is well suited to moving the excavated material.   
2.8. Autonomous Technology 
2.8.1. Power Supply 
The provisions of an autonomous vehicle power supply is an on going challenge 
in optimizing the balance between the amount of power required to function and the 
sustainability of the battery to allow it to run for long periods.  Unless they are tethered to 
an external power source, self-powered vehicles are limited in the amount of power 
  21  
carried by the overall size of the vehicle and the determining size of miniature robots is 
limited by the size of the power source [52].  As Sandia National Laboratories researcher 
Doug Adkins observed, the ?batteries need to run longer and smaller? [52].  Three watch 
batteries are used to drive his 41 cubic millimeter autonomous self-powered robot.  ?Low 
power processors and electronic tricks? helped Francesco Mondada and his research team 
at the Swiss Federal Institute of Technology in Lausanne, Switzerland to achieve the 
necessary power requirements to power their SWARM-BOT [53].  Like Doug Adkins? 
problems with miniaturization, complex mechanics and powerful electronics rapidly 
drain the energy available to propel the device.  The SWARM-BOT uses two lithium-ion 
cells to power the vehicle in standard operation; however power is saved by 
programming the processors to hibernate when not in use.  A group headed by J.M. Kahn 
assembled a smaller package by developing a self-powered vehicle through MEMS 
technology.  Called Smart Dust, this team from U.C. Berkeley is developing a vehicle 
that has the ability to sense and communicate, powered by solar cells, all in the space of a 
cubic millimeter [54].  In an attempt to avoid power consumption from on-board power 
supplies completely, some research groups are powering individual mechanisms and/or 
functions with separate power supplies [52].  It is suggested that the latter route should be 
followed with a digging vehicle.  The use of two power supplies would help to maximize 
the efficiency; with a deep draw cycle for powering the vehicle during digging and a light 
duty cycle after the final depth is reached and surveillance and/or communication is 
necessary. 
 
 
  22  
2.8.2. Single Vehicle vs. Group Approaches 
Research has expanded from developing a single autonomous vehicle able to 
handle multiple tasks to designing a range of autonomous vehicles able to perform 
specific tasks individually or as a group effort [53,55-60].  Robustness, reliability, 
efficiency, all-terrain capability, flexibility and adaptability are all important when 
building autonomous vehicles [53,56-60].  One of the ways to incorporate these qualities 
is to develop a group of vehicles that act as a unit, and this strategy has been titled 
?Swarm? technology [60].  From planetary exploration [52] to the ability to connect 
together as a single unit [53], the inspiration for the concept has come from the many 
social organizations found in Mother Nature.   
There are many advantages to swarm technology, with the biggest being the 
ability to continue a mission in spite of the loss of one or more of the members of the 
swarm, characterized as robustness.  In a single robot, programmed for many tasks, if any 
part of the robot fails, the mission correlating to that function has been lost.  However, in 
a swarm, if the same scenario occurs, another robot can complete the mission.  Currently, 
there is much discussion as to whether it is better to develop many simple separate tasked 
robots [53,56,58,60] or to create many robots that all have the same capabilities [54,61].  
In the case of exploration, it would be logical to disperse many vehicles that are capable 
of the same functions [55,61], especially when dealing with power supply issues.  The 
larger the area covered in a given time, the better.  Another benefit is the decrease in cost 
when mass-producing multiple robots.  Instead of developing a high-tech, one-of-a-kind 
robot, it may be more cost effective to produce many simpler robots.  The question then 
is, how do the members of the swarm communicate with each other? 
  23  
2.8.3. Communication 
Swarm communication has been inspired by many fields, but primarily ethology, 
or the study of animal behavior.  From the pheromones produced by bees and bee dances 
[62] to the indirect communications observed with ants [58,63], the ability to 
communicate in some manner is necessary or else efficiency is decreased [55,60,63].  
There are many possibilities when trying to develop communication within a swarm.  
One option is to use centralized control, with a main robot controlling others, while the 
main alternative is decentralized control, with each robot able to govern itself based on its 
programming and/or communication throughout the swarm.  However, centralized 
control has been proven to lack robustness.  If the main unit fails in any manner, then the 
rest of the swarm ceases to function due to a lack of orders.  Information received by the 
main unit can also become excessive with an increase in the number of units in the 
swarm.  As each robot transmits information to the main unit, then a bottleneck can 
occur, causing some information to be fragmented and lost altogether.  A more flexible 
and robust method is to decentralize the group, although there are various options within 
this scenario.  One question is whether the robots talk to the entire swarm (global) or to 
specific neighbors (local)?  Another possibility might be not to communicate at all, but 
simply to proceed based on the prevailing environmental conditions.  Kube and 
Bonabeau [58] successfully attempted this for a group of autonomous vehicles 
cooperating in a specific assignment of moving a box.  Based on the implicit 
communications of ants, the robots can sense the forces upon the box and decide whether 
it is helping or hindering the effort.  If it is helping, it continues with the process, 
  24  
however if a robot realizes that it is hindering the effort of the group then it will 
reposition itself until deemed helpful, by its own judgment.   
  25  
3. RESEARCH OBJECTIVES 
The original objective of this project was to design a vehicle that could dig 
through soil using a mechanism based on the mechanical action of Shape Memory Alloys 
(SMAs).  However, the small size of the SMA springs required many springs to be used, 
both in parallel (for force multiplication) and series (for stroke multiplication) with each 
other, in order to generate sufficient force and stroke.  The time needed for cooling 
between cycles was also a major issue.  Increasing the size of the springs might lead to 
needing fewer springs and having the ability to move at greater lengths per 
heating/cooling cycle, but would also increase the time needed to reach the austenite 
temperature and the time to cool back down for a repetitive cycle.  Testing revealed that 
larger springs suffered from ?memory loss? (what is sometimes referred to as ?shape 
amnesia?) at high currents, resulting in permanent deformations of the springs and they 
were thus deemed not suitable for this application. 
The project then moved on to consider the use of augers in conjunction with a 
small motor to create a digging vehicle.  Various sizes and shapes of augers were 
researched and tested.  The limiting factors for the design were found to be the size of the 
motor, the size of the auger, the nature of the coupling device and the amount of power 
that could be transferred to the auger.  Large augers created too much force for the motor 
to overcome, halting the descent, while smaller helical augers suffered from problems 
due to the devices used to couple them to the motor.  An auger designed specifically for
  26  
digging in sand proved too large for the available motor to handle.  As a result of the 
design flaws and failures experienced during the tests of shape memory alloys and 
augers, it became clear that a new approach is needed.  A look at the results of nature?s 
design process through examining the techniques used by fossorial creatures may inspire 
new designs for digging vehicles.  The preliminary study performed for this thesis may 
provide the basis for the selection of a creature for further study as a possible basis for a 
radically new strategic and scientific approach to developing a digging vehicle.
  27  
4. MATERIALS AND METHODS 
4.1. SMA Testing Procedure 
In previous years, tests performed by Dr. Daniel Butts and Dr. William Gale 
demonstrated the effective use of SMA springs as actuators for digging devices.  
However, due to the small size of these springs, many were needed in series and parallel 
to move the digging device.   The effort to manufacture springs that can produce a larger 
force than the existing prototypes required collaboration with an outside source, Shape 
Memory Applications, Inc.  The new springs were then characterized in terms of their 
stroke, time, force and current, over a range of values using a structure erected from 
metal tubing and anchored to a testing bench, as shown in Figure 2. 
  28  
 
Figure 2.  Testing platform. 
 
The test structure consisted of an upper arm, to which a SMA spring with a 
weight was attached, and a lower arm, to prevent the spring from overextending.  The 
spacing between the upper and lower arms was sufficient to allow the SMA spring to be 
stretched effectively, without harm, and could be adjusted depending on the size of the 
spring.  One end of the spring was attached by clamping the wire tightly between two 
washers on the upper arm, with the opposite end of the spring attached using the same 
method to a weighted hanger.  Calibrated cast iron weights, known as counterpoise 
weights, could be added to or removed from the hanger as needed.  Once the spring was 
  29  
stretched to the desired length, the resulting stroke distance was measured.  Two leads 
were placed at each end of the SMA with crimp connectors.  The wire was then 
resistance heated by using a DC power supply capable of variable voltage or current.  The 
power supply was attached to the spring by using wire clamps.   
 The first batch of larger SMA wires were shape set annealed with an oxide 
surface, with a nominal A
f
 (Austenite Finish temperature) of 46.2 ?C, a wire diameter of 
1.17 mm (~0.05 in), a total of 27 coils and a total length of 1.08 m.  Multiple tests were 
conducted with a fixed current and weight, recording the stroke and time for each.  The 
current ranged from one to six amps and the weights ranged from no additional weight 
(weight of spring) to 22.25 N (5 lbs).  The second batch of SMA wires had a larger 
diameter, measuring 1.75 mm (0.07 in).  Because these were significantly larger, the 
fixed current and weight were increased.  Testing started at a 44N (10 lb) load at 10 A 
and increased in increments of 8.9 N (2 lbs) and 2 A up to 89 N (20 lbs) and 20 A.  An 
interval of 300 s (5 min.) was allowed between each individual test, to permit sufficient 
cooling of the relatively heavy gauge SMA wire.   
 The stroke length was determined by reading the length from the yard stick.  
Before the current was applied, the yard stick was aligned with the lower end of the 
spring (i.e. the end with the weight attached).  After the current was applied and the SMA 
returned to its original shape, the resulting length was recorded.  The time for each test 
was recorded using a digital stopwatch.   
  30  
4.2. Auger Testing 
4.2.1. Electric Motor 
To provide a comparison for the SMA results, digging tests using a direct current 
(DC) electric motor were conducted.  The research team, as a sufficient starting depth, 
agreed upon a 209 L (55-gallon) drum with a depth of 0.90 m (36 in) and diameter of 
0.58 m (23 in).  Off the shelf silica sand, dried and non-compacted, was used for the 
testing bed.   
Testing consisted of digging in the 55-gallon drum at various fixed voltages, 
recording the depth reached and time for each.  The power supply used in the SMA 
testing was also used to power the motor.  The depth achieved was expressed in terms of 
the amount of vehicle buried rather than an absolute measured depth.  The time was 
recorded using a stopwatch.   
4.2.2. Conical Auger 
Initial tests were conducted using a conical auger (see Appendix A for 
dimensions) designed by Rob Langley at Auburn University and fabricated by the 
sponsor.  An adapter, designed by Daniel Butts at Auburn University and supplied by the 
sponsor, was used to attach the motor (also supplied by the sponsor) to the conical auger.  
As shown in Figure 3, the adaptor acted as a reducer to accommodate the large diameter 
of the auger compared to that of the motor.  The plate was secured to the auger and motor 
with bolts and a setscrew, respectively.   
 
  31  
 
Figure 3.  View of machined attachment connecting the conical auger to the motor. 
 
 
The digging tests used fixed voltages of 5, 10, 15 V with time and current being 
recorded at auger depths of quarter, half, three-quarter, and fully buried, as well as the 
final point of descent (i.e. maximum attainable depth).  Each test was run three times, 
giving a total of nine digging trials.   
To keep the motor stationary, different anti-counter rotation devices such as large 
flat fins and cylindrical rods were tested.  However this problem was deemed secondary 
compared to the overall goal enhanced digging ability, as an anti-counter rotation device 
could be retrofitted once a mechanism was selected.  Different fin designs that were used 
  32  
for anti-counter rotation can be seen in Appendix B.  Due to the short duration of the 
project, the notion of designing an optimal fin design was abandoned.  Two metal strips, 
coaxial with the motor, were attached using three metal clamping bands.  The metal strips 
acted as the stability mechanism for a 0.64 m (24 in) long metal bar, which was placed 
over the strips perpendicular to the motor.  During operation, the bar was steadied by 
hand to prevent the motor from counter rotating (Figure 4).   
 
 
Figure 4.  View of motor with attached rod for counter rotation. 
  33  
4.2.3. Helical Augers 
As a comparison with the conical auger, ordinary helical, gardening augers were 
used.  These augers varied in diameter, pitch and overall length, with all tests conducted 
under the same conditions.  Initial tests used off-the-shelf, one-piece clamp-on couplings 
with socket-head cap screws (Figure 5).  However these were not sufficient, as the overall 
forces needed to turn the auger were greater than the coupling device could handle and 
led to the auger shaft turning within the bore hole of the coupling device.  
 
 
Figure 5.  View of the coupling devices connecting the helical augers to the motor. 
 
 This observation led to the design of a custom attachment piece constructed from 
an aluminum block.  The attachment piece was designed with a conical shape to help in 
the flow of the material from the top of the auger to the motor body.  The motor and 
  34  
auger were able to slide in on opposite ends, with set screws holding each in place 
(Figure 6).  If a simple cylindrical piece, matching the diameter of the motor, was 
manufactured, it is believed that material flow would be hindered between the auger and 
motor, creating unnecessary forces that must be overcome and hence stripping of the 
auger in the coupling devices. A thin piece of Teflon was placed between the motor and 
attachment to help reduce the frictional forces between the two.   
 Helical augers used in the digging tests consisted of 64 mm (2.5 in) and 76 mm (3 
in) diameter by 200 mm (8 in), purchased from A.M. Leonard, Inc.   
 
 
Figure 6.  View of adaptor connecting the helical augers to the motor. 
  35  
4.2.4. Torque Requirements 
To achieve the force needed for the vehicle to dig into sand, different tests were 
conducted depending upon the type of forces the vehicle would encounter.  Before an 
optimal auger could be designed, a determination of the amount of torque required for 
various augers to dig effectively was needed.  Three groups were tested: an extremely 
large helical auger (200 mm, 8.0 in diameter), similar diameter conical augers (127 mm, 
5.0 in and 140 mm, 5.5 in) and similar diameter helical augers (64 mm, 2.5 in and 76 
mm, 3.0 in).  A photograph of these five augers can be found in Figure 7. 
 
 
Figure 7.  View of the augers tested. 
  36  
 For the conical augers, a quick adaptor was installed to measure the torque 
required to turn the auger at different depths.  The adaptor was held in place with bolts, 
on top of the auger, and a ? in nut was firmly set in the middle of the attachment.  A 
hand-held torque wrench was coupled with the auger, with an extension placed on the 
nut. 
4.2.5. Push Test 
To gain a better understanding of the frictional forces the vehicle was likely to 
encounter from contact with its body (during the present project the body was generally 
only that of the motor) with the surrounding sand, a ?push test? was needed, in which the 
force needed to push the body into the sand was measured.  Initially, a Teflon rod ~ 51 
mm (2 in) diameter, with a cone tip, was used for the tests in order to quickly estimate the 
force required to overcome frictional drag on a similar size vehicle body in a sand 
medium.  The force measured increased exponentially with increasing depth.  It was 
observed that with the same rod in identical tests, the forces experienced were not 
consistent from run to run, but instead increased with each run into the same sand. 
It should be noted that the depth of the bucket used was only 381 mm (15 in) and 
the exponential increase in force could be, at least in part, a result of this factor.  Also, it 
is believed that the increase in force with each run was due to the increase in sand 
compaction resulting from the previous run.  With these considerations in mind, a new 
method of experimentation then evolved.  To help eliminate the variables mentioned, a 
new container was employed with a depth of 610 mm (24 in), as compared with only 381 
mm (15 in) for the previous setup.  The diameter of the container (356 mm, 14 in) was 
chosen to better simulate a semi-infinite medium, in an attempt to minimize the artificial 
  37  
forces on the sides of the vehicle body created by the constraint of the sand interface with 
the container.  After each trial run, the sand was removed from the apparatus and gently 
?poured? back into the container, letting the sand particles randomly settle.   
The test rods were comprised of five aluminum rods, each 300 mm (12 in) length, 
with outside diameters (OD) that ranged from 25 to 76 mm (1 to 3 in), in increments of 
~13 mm (.5 in) (Figure 8).   
 
 
Figure 8.  Aluminum rods used for the push test. 
 
Each rod had a conical tip, with a 75? pitch.  A female thread was machined on 
the opposite end of the rod, for use in securing the test rod to the load cell.  The latter was 
needed to help maintain a consistent insertion point, normal to the surface of the sand, as 
a major misalignment would cause a slanted insertion and would therefore have produced 
an increase in the force needed, skewing the results.  Each test was run at a speed of 75 
mm/min (~3 in/min) to a final depth of 500 mm (~ 20 in) with a 10 Hz data acquisition 
rate on a MTS tensile machine (Mechanical Testing and Simulation Inc., www.mts.com). 
  38  
4.2.6. Pull Tests 
An experiment was performed to determine how much force an auger can pull.  
Using the 64 mm (2.5 in) diameter auger, a pulley system was constructed to allow the 
auger to pull various weights (Figure 9).   An elongated auger was added (Appendix C) to 
isolate the digging strength of the auger from the frictional forces exerted by the motor 
body.  Using a similar setup as previously used for the SMA testing, two pulleys, spaced 
150 mm (6 in) apart, were added to the top of the structure.  Extension bars were added to 
the motor body (Figure 9) to allow one end of a thin wire to be attached.  The wire was 
then pulled through the pulleys using varying weights.   
 
  39  
 
Figure 9.  Pull test pulley system. 
4.2.7. Colored Sand Experiment 
A container of layered, colored sand was used to reveal how the sand moved as a 
fluid body around the conical auger and motor.  The set up used in this experiment can be 
seen in Figure 10, with five different colors layered in approximately 75 mm (~3 in) 
increments.  Three digging experiments were run and documented using a digital video 
recorder.   
  40  
 
Figure 10.  Multi colored sand used to examine the auger-sand interactions. 
 
After the initial trials, an experiment using a ?sand freezing? process was 
conducted.  The use of sodium silicate is well established in foundries for binding sand 
cores [64].  With the appropriate amount of sodium silicate mixed with the sand, the 
application of carbon dioxide (CO
2
) then ?freezes? the mixture at any selected moment 
during the digging process.  The sodium silicate acts as a binder, keeping the sand mold 
intact without the need for any outside force.  This principle was used in the colored sand 
experiment, freezing the sand around the conical auger and motor and then slicing it open 
to observe the sand-auger interaction.  It was observed that the sand ?freezes? well when 
  41  
doubling the manufacturer?s suggested 3 ? 4 % by weight of sodium silicate; it was found 
that a mixture of 8 % by weight of sodium silicate kept the sand core stable enough to 
prevent crumbling when sliced into sections after CO
2
 exposure.  After mixing the sand 
with the sodium silicate manually, the colored layers were placed in the testing apparatus.  
The motor and conical auger were then placed in the center of the surface and allowed to 
dig.  Upon the auger reaching bottom, or stalling, a CO
2
 lance was inserted in various 
spots to allow the gas to diffuse throughout the sand.  After hardening had occurred, a 
quarter piece of the sand ?box? was cut out to show a profile of the sand with the auger 
and motor. 
4.2.8. Customized Auger 
As a late addition to the digging experiments, Dr. Jason Baird from the University 
of Missouri-Rolla, designed a custom auger specifically for digging in sand.  Having 
acquired the necessary testing equipment, Auburn University agreed to test the design 
provided by Dr. Baird and his staff.  Dr. Baird?s auger can be seen in figure 11.  A 
detailed AutoCAD drawing of the auger can be found in Appendix D.  The design is 
based on a literature review that Dr. Baird modified by incorporating a reverse flight at 
the top of the auger to keep the material from falling inward on the borehole and to keep 
a continuous force on the auger bit.  As the bottom flights are forcing material up, the 
reverse flight acts in the opposing direction to keep forcing the auger downward. 
 
  42  
 
Figure 11.  Digging auger designed by Dr. Baird, professor at the University of 
Missouri-Rolla.   
 
4.2.8.1.Plain Sand Experiments 
For comparison with the previous augers, tests were conducted in a 55-gallon 
drum filled with plain sand using the anti-counter rotation device shown in Figure 4.  An 
attachment with a 100 mm diameter and 50 mm (2 in) length was constructed.  Four bolts 
held the attachment in place with the auger, while the motor slid into a machined keyway 
in the center of the device. 
 The auger was tested using three different fixed voltages (5 V, 10 V or 15 V) with 
variable current.  Each voltage level was repeated three times, for a total of nine trials.   
The time and current at four auger depths, a quarter (130 mm, 5 in), half (230 mm, 9 in), 
  43  
three-quarters (380 mm, 15 in), and final auger movement, were recorded.  Before each 
trial, the sand was removed from the tank and reloaded in order to avoid compaction and 
maintain uniformity with earlier trials. 
4.2.8.2.Sand Mixture Experiments 
Personal communication with Dr. Baird has shown that pure sand is one of the 
most challenging materials to dig in, resulting in a stripping of the borehole walls, that 
may cause the auger to spin without descending further.  It was thus decided to create a 
custom mixture of sand, flour and oil with a ratio of 1:1:0.2 by weight, respectively.  To 
create this mix, off the shelf baking flour and technical grade mineral oil were mixed with 
the sand to fill a 114 L (30 gallon) drum with an OD of 0.41 m (16 in) and depth of 0.79 
m (31 in).  Mixed in small batches, the sand bed consisted of approximately 61 kg each 
of flour and sand, and 12 kg of mineral oil.  To avoid creating a sand gradient, each new 
batch was mixed in with the previous layer, as opposed to simply being laid on the top.  
This improved the homogeneity of the mixture from top to bottom.   
 The auger was tested using three different fixed voltages (5 V, 10 V or 15 V) with 
variable current.  Each voltage level was repeated three times, for a total of nine trials.  
The time and current at a series of auger depths of a quarter (130 mm, 5 in), half (230 
mm, 9 in), three-quarters (380 mm, 15 in), and full (460 mm, 18 in), followed by multiple 
motor depths of a quarter (560 mm, 22 in), half (620 mm, 24 in), three-quarters (660 mm, 
26 in) and final auger movement (~710 mm, 28 in), were recorded.  Before each trial, the 
mixture was removed from the tank and reloaded, letting it settle naturally, in order to 
prevent compaction. 
  44  
4.2.8.3.Compacted Sand Mixture Experiments 
After each trial, the sand mixture was removed and then immediately reloaded 
into the container.  However, the mixture at the bottom was inevitably compacted more 
than the mixture at the top due to the additional weight of the sand mixture.  This was 
evident when the vehicle was placed on top of the sand bed, where it began to descend 
into the sand mixture even before the motor was started.  In order to eliminate the effect 
of this compaction gradient, additional trials with a compacted sand bed were necessary, 
as the compacted bed would be more representative of a real soil and thus provide a 
better estimate of the required digging energy.   
The auger was tested using three different fixed voltages (5 V, 10 V or 15 V) with 
variable current.  Each voltage level was repeated three times for a total of nine trials.  
The time and current at multiple depths discussed previously were recorded.  Before each 
trial, the mixture was removed from the tank and reloaded, being allowed to settle 
naturally.  After a depth of 100-130 mm (4-5 in) had been poured in, a weight of 30 kg 
(66 lbs) was gently placed on the surface to compact the sand.  This process was repeated 
until the testing container was full.
  45  
4.2.8.4.Torque Requirements of Customized Auger Experiments 
A quick adaptor was installed in order to measure the torque required to turn the 
auger at different depths.  The adaptor was held in place with bolts, on top of the auger, 
and a ? in nut was firmly set in the middle of the attachment.  A hand-held torque 
wrench was coupled with the auger, using a ratchet extension placed on the nut. 
4.2.9. Digging Energy Requirement 
The digging energy (E
dig
) was estimated based on the volume of sand displaced 
(Equation 1, [65]), 
 
?
?
?
?
?
?
?
?
=
t
d
A
IV
E
dig
max
 (1) 
and based on the volume of auger buried (Equation 2, [65]),  
 
?
?
?
?
?
?
?
?
?
=
t
Vol
IV
E
auger
dig
max
 (2) 
where V is the applied voltage, I
max
 is the maximum current, A is the cross-sectional area, 
d is the depth, Vol
aug
 is the volume of the auger and t is the time. 
4.3. Vehicle Experimentation 
4.3.1. Earthworm Concept 
To aid in the digging process, a model based on the way an earthworm travels was 
tested. Rather than simply pushing the displaced sand around the vehicle body, an 
attempt was made to allow the sand to move through it.   
  46  
As shown in Figure 12, the motor and a 76 mm (3 in) auger were placed inside a 114 mm 
(~ 4.5 in) diameter tube, leaving about 89 mm (3.5 in) of the screw exposed below the 
tube housing to permit the auger to gain a  purchase in the sand. 
 
 
Figure 12.  Earthworm design; Motor and 76 mm auger within a tube. 
 
4.3.2. Friction Reduction 
Another method designed to aid digging that was tested was to reduce the surface 
friction experienced around the motor?s body.  Here, air was used as the friction-
reduction medium.  Small diameter reinforced tubing of various lengths was placed along 
the body of the motor and pulses of air blow through the ends of the tubes.  This caused 
air to blow along the surface, loosening up the sand particles and reducing the surface 
friction generated by the sand to motor interaction. 
  47  
4.4. Biomimetic Analysis 
4.4.1. Brainstorming 
The primary goal of this portion of the project was to generate an analogue for 
implementation in a prototype vehicle designed to dig to a desired depth.  The project 
was not expected to identify a suitable power source or provide, a global positioning 
system for tracking, the ability to steer, or any of the many other systems needed to create 
a functional vehicle.  However, after piecing together off the shelf products to try to 
achieve this goal, it became obvious that a different approach was needed.  An ?outside 
the box? view came through the inspiration of ?Mother Nature?.   
Instead of trying to re-invent the wheel, the researchers chose to take a closer look 
at creatures that dig successfully on their own.  A search of the relevant literature 
identified a host of creatures, including mammals, reptiles, insects and invertebrates, in 
numbers far too great to be useful in this study.  The goal was to model something that 
could dig and hence a ?weeding out? or selection criteria had to be stipulated. 
4.4.2. First Cut 
The first selection criteria used was the ability to dig to at least one body length in 
depth.  Many creatures that do not have an underground habitat do nonetheless dig 
routinely.  For example, dogs dig to bury or unbury things, to escape other predators or 
just out of apparent boredom.  The fact that creatures dig is not sufficient grounds to 
justify an attempt to emulate their actions; therefore the ability to dig deep enough to 
cover one?s self completely was selected as a criterion.  This criterion was fully satisfied 
with creatures that are characterized as ?fossorial?, meaning that they live underground.  
Those creatures not characterized as fossorial were eliminated from this study.  The 
  48  
faculty and students at the University of Michigan have generated an online database of 
wildlife for public use, and this was used extensively to provide the central core of the 
project?s research.   
This website, animaldiversity.ummz.umich.edu, is an online encyclopedia of 
animals to which students at the University of Michigan contribute the most up to date 
information gained by researchers.   By entering ?fossorial? into the animal diversity?s 
search engine, the first list of animals was selected. 
4.4.3. Initial Round of Questions 
The initial selection produced a list of about 50 species of animals that are 
fossorial.  In order to avoid documenting the life history of each animal, a list of general 
questions (Appendix E) was used to select the pertinent details.  The website only 
contains brief descriptions about the animals and not all the necessary information could 
be found elsewhere.  Not all animals have been studied to the same extent, as some are 
observed far more intensively than others.  Irrespective of whether the reason for the 
study is economic, academic, or maybe environmental, some animals have not been 
studied in enough detail to provide all the necessary information.  However, by seeking to 
answer specific questions for each, the differences between the animals became more 
apparent and more focused questions could be included in the next round. 
4.4.4. Second Round of Questions 
After learning more about the differences between the animals, a second list of 
questions was used to gain specific knowledge.  These questions generally examined 
issues related to the metrics used to describe the animals, which were divided into three 
main categories: general characteristics, digging mechanism and bioefficiency.   
  49  
 The general characteristics described the appearance and the social aspects of the 
creatures.  The second category, digging mechanism, comprised four subcategories:  
performance, namely the depth and rate of the digging; mechanical, or details of the 
digging apparatus of the creature, such as motion and efficiency; chemical, as some 
creatures have chemical capabilities to aid in their digging process; and environmental, to 
examine the conditions and the materials they dig in.  The third and final category is 
bioefficiency, which considers the availability and calorific value of the food sources 
ingested and how they give the individual creatures the energy they need to dig. 
4.4.5. Weighting System 
Not all this information was readily available, however this did provide a way to 
separate the creatures from one another, and compare the best candidate versus the worst 
candidate.  After reviewing the possible metrics, a new categorization was established, 
that focused on the information that would be important for replicating the creature:  
speed, depth, power, material substance, mechanical complexity, geographic range, and 
the size of the body of the creature.  This new set of categories was used to create a 
weighting system that could be used to assign a numerical value for each creature.   
 Each category had its own weighted percentage and set of related metrics.  Each 
metric was rated on a scale of one to five, based on the information known.  However 
important the metric might be, not all of the information was available.  The lack of 
information could not be used to justify rating a metric as unnecessary, as this metric 
might still be vital.  For this reason, two calculation methodologies were considered, both 
with and without an allowance for unknown information.  Each metric had two values: a 
raw score using arbitrary values (e.g. 3 out of 5) for the unknowns, and a raw score 
  50  
implementing only the known values.  For organisms with little available data, the use of 
an arbitrary value for unknown categories tended to push the score towards an ?average? 
score, whereas ignoring the categories without data tended to overweight the importance 
of the available data.  Using both methods did not correct for this shortcoming, but it did 
help in interpreting the meaning of the scores. 
 After all the metrics had been assigned an appropriate score, the raw scores were 
weighted and added together for a final score.  Each creature had two final scores; with 
and without the arbitrary values for the unknown.  Because the digging vehicle would be 
used in many different situations, each creature was evaluated for three desired scenarios:  
Scenario 1, small body (? 300 mm; 12 in) at low depths (? 1.52 m; 60 in); scenario 2, 
small body (? 300 mm; 12 in) at medium depths (1.52 m; 60 in ? 6.10 m; 240 in) and 
scenario 3, medium body (300 mm; 12 in ? 910 mm; 36 in) at deep depths (? 6.10 m; 240 
in).  Thus, each creature had six final scores, two for each scenario (Appendix F). 
4.4.6. Flow Chart for Biomimetic Design 
After the creature has been chosen based on the final scores, an overall look was 
needed for design purposes.  A flow chart was created to aid in this process.  The flow 
chart was designed to identify the power requirements, materials and any special features 
that need to be incorporated into the final vehicle and the overall cost effectiveness of the 
vehicle.  After all these aspects have been covered, if it was found that the chosen 
creature was not cost effective, a redesign would be necessary in order to select cheaper 
materials, an alternative power source or possibly a different creature that would be more 
suitable overall.  This process can be seen in Appendix G. 
  51  
5. EXPERIMENTAL RESULTS 
5.1. SMA Testing Results 
5.1.1. First Batch 
The springs were elongated to their maximum length (without resulting in 
irreversible damage) of about 305 mm (12 in).  While applying a constant current and 
load, the actuation distance was measured manually, hence human error, (approximately 
+/- 1 mm), may account for a portion of the variability observed between tests of similar 
current and load.  Table 1 shows the observations for springs at different loads and 
currents.  It should be noted that all tests were performed in constant current mode and 
the voltage was allowed to fluctuate. This is important since a change in resistivity is 
observed when the wire undergoes its phase transformation.  The voltages noted are the 
averages seen after a drop, due to the lower resistance once actuation has occurred. 
 The average electromagnetic actuation efficiency of the SMAs was calculated 
using the following expression [20]: 
Eff
av
 = 
(sec)
)()(
timeampsvolts
mStrokeNForce
??
?
    (3) 
 
which corresponds to the mechanical work divided by the electrical energy supplied.
  52  
Table 1.  SMA conversion efficiency of the first batch of springs  
Force Other Parameters 
Lbs N 
Stroke 
(m) EMF ( V) 
Current 
(A) 
Time 
(s) 
% 
Efficiency 
0.02 0.07 0.17 2.5 3 49 0.003 
1 4 0.13 2.5 3 59 0.13 
2 9 0.07 2.5 3 75 0.11 
3 13 0.01 2.5 3 87 0.03 
4 18 0 2.5 3     
5 22 0 2.5 3     
              
Force Other Parameters 
Lbs N 
Stroke 
(m) EMF ( V) 
Current 
(A) 
Time 
(s) 
% 
Efficiency 
0.02 0.07 0.16 3.3 4 25 0.003 
1 4 0.13 3.3 4 30 0.15 
2 9 0.11 3.3 4 39 0.19 
3 13 0.08 3.3 4 45 0.17 
4 18 0.05 3.3 4 67 0.10 
5 22 0.01 3.3 4 94 0.02 
              
Force Other Parameters 
Lbs N 
Stroke 
(m) EMF ( V) 
Current 
(A) 
Time 
(s) 
% 
Efficiency 
0.02 0.07 0.15 3.7 5 19 0.003 
1 4 0.13 3.7 5 25 0.12 
2 9 0.10 3.7 5 30 0.17 
3 13 0.08 3.7 5 42 0.14 
4 18 0.05 3.7 5 42 0.12 
5 22 0.03 3.7 5 70 0.05 
              
Force Other Parameters 
Lbs N 
Stroke 
(m) EMF ( V) 
Current 
(A) 
Time 
(s) 
% 
Efficiency 
0.02 0.07 0.16 4.7 6 12 0.003 
1 4 0.14 4.7 6 15 0.14 
2 9 0.11 4.7 6 20 0.17 
3 13 0.09 4.7 6 31 0.14 
4 18 0.07 4.7 6 39 0.11 
5 22 0.04 4.7 6 44 0.08 
 
 
  53  
To better visualize the capability of the SMAs, a graph of current vs. stroke was 
generated from these results.  Figure 13 represents the average maximum SMA spring 
contraction distance (i.e. stroke) obtained when lifting various weights (0-22 N in 4 N 
increments, 0-5 lbs in 1 lb increments) versus a range of input currents (3-6 A in 1 A 
increments).  Three amperes was the minimum current tested due to the fact that anything 
less provided insufficient heating of the SMA spring.  Six amperes was the maximum 
current tested since this was the largest integer value that the power supply employed at 
that time was capable of producing.  The graph shows that large strokes can be obtained 
at low or no loads with minimal energy input (e.g. 3 A at 2.5 V = 7.5 W).  However, at 
greater loads (i.e. 13-22 N, 3-5 lbs) more energy input is necessary (e.g. 5 A at 3.7 V = 
18.5 W).  In addition, considerably smaller strokes were achieved at greater loads.  This 
is understandable, considering that SMAs can produce either large strokes or large forces, 
but not both simultaneously.  Note that the error bars included in the graph show the 
variability in the strokes obtained.  This variability is attributed to both multiple 
individuals performing the tests and intrinsic variability from spring to spring. 
  54  
Current vs Stroke
0
20
40
60
80
100
120
140
160
180
200
3456
Current (A)
St
r
o
k
e
 (
mm)
0 N (0 lb)
4 N (1 lb)
9 N (2 lb)
13 N (3 lb)
18 N (4 lb)
22 N (5 lb)
 
Figure 13.  Graphical representation of the current vs. stroke results with error bars 
showing the standard deviation. 
 
Figure 14 represents the average time needed to actuate when subjected to various 
currents versus the same amount of weight.  It should be noted that at 3 A and above a 
force of 13 N, the time required for actuation is much greater when compared to those at 
higher currents.  
 
  55  
Force vs Time
0
10
20
30
40
50
60
70
80
90
100
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Force (lbs)
Ti
m
e
 (s)
3 A
4 A
5 A
6 A
 
Figure 14.  Graphical representation of force vs. time results (note at 3 A lifting of a 3 N 
load did not occur so the time is not shown). 
 
5.1.2. Second Batch 
It was assumed that the larger springs would be able to handle larger loads along 
with a larger current, so testing was to start at a 45 N (10 lbs) load at 10 A, increase in 
increments of 9 N (2 lbs) and 2 A up to 89 N (20 lbs) and 20 A with a 300 second (5 
minute) cooling period in between each run.  Three springs were tested, however, the 
actuation distances were not repeatable.  Individual spring results are in Appendix H.  
During one run, after about 20 s at 20 A, deformation of the spring occurred as the 
current was being applied.  The spring literally started to uncoil and stretch with the 
weight still attached.  The attainable actuation distance dropped from 170 mm (initial 
test) to 90 mm (after deformation).  This same abnormality was observed even at lower 
currents and loads.  Table 2 shows the results from a spring at these lower conditions.  
  56  
The actuation distances in the second round of testing at the same (low) currents were 
reduced when compared to the initial distances.  These observations are the same as seen 
in previous experiments; however, the differences were not as extreme as previous 
recordings.  Thus it might be that the lower currents were still harmful but took longer to 
cause extreme damage. 
 
Table 2a. Initial observations of a SMA spring at low currents and loads. 
Force Other Parameters 
Lbs N 
Stroke 
(m) EMF ( V) 
Current 
(A) 
Time 
(s) % Efficiency 
6 27 0.19 3.2 9 34 0.52 
6 27 0.19 3.2 9 32 0.56 
6 27 0.18 3.2 9 33 0.51 
6 27 0.17 2.9 8 43 0.46 
6 27 0.17 2.9 8 44 0.45 
6 27 0.17 2.9 8 43 0.46 
6 27 0.17 2.5 7 107 0.25 
6 27 0.16 2.5 7 109 0.23 
6 27 0.16 2.5 7 107 0.23 
 
2b.  Results of SMA spring at low currents and loads; second testing unable to achieve 
initial results. 
Force Other Parameters 
Lbs N 
Stroke 
(m) EMF ( V) 
Current 
(A) 
Time 
(s) % Efficiency 
6 27 0.17 3.2 9 36 0.44 
6 27 0.17 3.2 9 35 0.46 
6 27 0.17 3.2 9 33 0.48 
6 27 0.16 2.9 8 44 0.42 
6 27 0.16 2.9 8 45 0.41 
6 27 0.16 2.9 8 44 0.42 
6 27 0.17 2.5 7 106 0.25 
6 27 0.16 2.5 7 105 0.24 
6 27 0.15 2.5 7 107 0.22 
 
 
The next point of interest was the SMA actuation efficiency, expressed in terms of 
converting electric power to useful work.  It was found that the range of efficiency was 
very low for the first batch, with the efficiency ranging from as low as .003% to as high 
  57  
as 0.18%, depending on the amount of weight and/or current (Table 1).  An average 
energy conversion efficiency for a ?typical? NiTi SMA is only about 5%, as most of the 
energy is dissipated in heating the wire before the transformation begins, especially given 
the high heat loss from the large surface area of a coiled spring.  Nonetheless, the results 
were still far below the ?accepted? efficiency of Nitinol SMAs.  One possible explanation 
of the low efficiency was that the heating cycled employed for the work was not 
optimized for the wires used. It was theorized by the project team that higher currents, 
such as 9-10 A, should decrease the amount of time needed for the wire to go through 
transformation.  An optimum compromise between the two variables was needed. 
In contrast to the small energy conversion efficiency, the wire had a very high 
lifting capability to spring weight ratio.  With a total weight of 7 g, the SMA wire can lift 
22 N, giving it a 3 N per gram (~ 3.0 lbs/oz) ratio. 
 It was noted that the efficiency of the second half of the SMA order had greatly 
increased over that of the first half.  In the case of spring 1, the overall average of the 
efficiency was 0.5 % ? three and half times that of the first batch efficiency estimate of 
0.14 %.  In the case of spring 2, the efficiency dropped from 0.5 % to 0.4 %.  In the case 
of spring 3, initial testing showed an average of 0.46 % and a drop to 0.41 % for the 
second testing. 
  58  
5.2. Auger Testing 
5.2.1. Torque Requirements 
Figure 15 shows the amount of torque needed to dig as the depth increases.  All 
three groups of augers required increased torque with depth, although some experienced 
this increase at a higher rate than others.  Using a hand held torque wrench, the data may 
be somewhat variable due to human error (approximately +/- 5 Nm).  Overall, it was 
observed that the torque requirements needed to turn the large 200 mm (8 in) auger were 
more than that was physically possible for the experimenters below 305 mm (12 in) of 
sand.  Conical augers of similar diameter experienced a steady increase in torque 
throughout the testing, but stopped at ~ 610 mm (24 in) due to equipment limitations (i.e. 
exceeded the range of the torque wrench).  The smaller diameter set of similar helical 
augers experienced the least amount of torque needed when compared at the final depth 
of 610 mm (24 in).   
 
 
  59  
Rough Depth vs Required Torqe for Several Auger Types
0
20
40
60
80
100
120
140
160
180
200 250 300 350 400 450 500 550 600 650
Depth (mm)
Tor
q
ue
 (
N
m
)
0
200
400
600
800
1000
1200
To
r
que
 (
i
n-
l
bs
)
63 mm (2.5 in) Helical Auger
76 mm (3 in) Helical Auger
203 mm (8 in) Helical Auger
127 mm (5 in) Conical Auger
140 mm (5.5 in) Conical Auger
 
Figure 15.  Depth vs Torque for various auger diameters. 
 
5.2.2. Push Test  
The results were generally as expected, with the required force greatly increasing 
with increasing rod diameter (Figure 16).  As the depth increased, so did the amount of 
force needed to compensate for the increasing contact area with the sand and, hence, to 
overcome the frictional forces on the rod.  At about 300 mm (12 in) depth, the rod was 
fully inserted and the entire rod surface area was in contact with the surrounding sand, yet 
there was still a significant increase in the amount of force needed to continue to push in 
the larger diameter bars.  Some of this could be explained by the use of a push rod which 
attached the push test sample to the screw driven testing machine (Mechanical Testing 
and Simulation Inc., www.mts.com), but the diameter of this push rod was only ~25 mm 
(1 in) and so this should have been relatively insignificant when pushing the ~51 mm (~ 2 
  60  
in) and larger OD samples.  The observation that the required insertion force increased 
with depth beyond the sample length for the large diameter rods (50 mm, 2 in and greater 
diameter), but leveled off somewhat for smaller diameter rods, has yet to be explained. 
 
Averages of Al Rods
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
0 50 100 150 200 250 300 350 400 450 500 550
Distance (mm)
For
c
e
 (
N
)
0
250
500
750
1000
1250
1500
1750
2000
2250
2500
2750
3000
3250
3500
Fo
r
c
e
 (
l
b
f
)
25 mm (1 in)    OD Al Rod
38 mm (1.5 in) OD Al Rod
51 mm (2 in)    OD Al Rod
64 mm (2.5 in) OD Al Rod
76 mm (3 in)    OD Al Rod
 
Figure 16.  The average forces needed to push the different OD aluminum rods into a 
sand medium at 75 mm/min (3 in/min).   
 
5.2.3. Pull Test 
Using the 64 mm (2.5 in) diameter auger, it was found that the auger with the 
motor, which weighed 46 N (10.5 lbs), could only pull a load of 55 N (12.3 lbs) before it 
stopped digging.  With the auger ~ 250 mm (10 in) into the sand, weight was added and 
the motor was observed to see if any further descent occurred.   
 
  61  
With the addition of 4 N (1 lb) to the 55 N, the auger ceased to bite and just spun in the 
sand.  Subtracting the weight of the motor from the weight pulled, the total force was 
about 8 N. 
 The experiment was repeated with the ~ 75 mm (3 in) auger, however an 
elongated auger was not available at the time and so testing proceeded with the normal 
set up as shown previously in figure 6.  The results showed a considerably different 
behavior, however; with weights up to 46 N (10.5 lbs) the auger was able to dig to about 
480 mm (19 in) deep, which is the length of the auger and motor body combined.  Not 
only was the auger pulling the weight, but it was also overcoming the frictional forces 
encountered due to the surface of the motor.  The only hypothesis to explain this 
observation was that since the auger had been severely roughened from the previous 
digging trials in the sand it was able to bite the sand in a more effective manner.  The 
increase in diameter might also have aided in the digging, giving the auger more surface 
area to interact with the sand. 
5.2.4. Colored Sand Experiment 
The profile seen in Figure 17 showed little sand mixing, as the layers were still 
largely intact around the auger.  Teeth marks from the auger can also be seen from the 
downward motion.  The dip in the colors in each layer that is visible to the right of the 
auger was artificially created and should be excluded from any observations.  This was 
due to the insertion of the lance used to add the carbon dioxide.   
 
  62  
 
Figure 17.  Profile of digging trial using a sand core. 
 
Observations made with the frozen sand confirmed that sand was not carried up to 
the surface.  This indicates the auger was pushing the displaced sand outward, rather than 
carrying is upward, as an auger should.  Indeed, the walls of the sand tank were observed 
to bulge outwards during digging, as would be expected if sand was displaced outwards, 
rather than upwards.  This would seem, at least in part, to explain the large torque needed 
for digging.  An additional observation was that the tip of the auger appeared to have 
churned up the sand, rather than gripping it, in a fashion reminiscent of a screw that has 
stripped its hole.  Thus, the auger employed for the test appeared to have lost its grip on 
the sand.  This would account for the observation that the digging tended to stall, before 
the torque capabilities of the motor were exceeded. 
Undisturbed layers 
Auger Teeth 
Lance Insertion 
  63  
5.3. Customized Auger 
5.3.1. Plain Sand  
The results, average time, average current, standard deviation (STDEV) for time 
and current and estimated digging energy are summarized in Table 3a and 3b.  As in 
previous experiments, the digging energy has been calculated based on both the volume 
of sand displaced and the depth to which the auger is buried (see 4.2.9 Digging Energy 
Requirement, Equation 1 & 2) 
 
Table 3a.  Digging results using fixed voltages of 5, 10 and 15 V.  Average values for 
three trials. 
 
 
Begin Digging Auger Quarter 
Buried 
Auger Half 
Buried 
Auger Three
-
Quarters Bu
ried 
Final 
Average Time (s) 0 61 160 410 440
STDEV (Time) 0.0 11 14 27 5.0
Average Amperes (A) 0.6 1.1 2.0 5.4 6.2
5 V  
STDEV (Amperes) 0.0 0.02 0.10 0.3 0.01
  
            
Average Time (s) 0 31 71 145 170
STDEV (Time) 0.0 3.1 3.2 6.0 3.0
Average Amperes (A) 0.6 1.1 1.6 4.6 6.2
10 V 
STDEV (Amperes) 0.0 0.01 0.4 0.5 0.01
              
Average Time (s) 0 20 47 89 110
STDEV (Time) 0.0 1.0 3.2 5.5 2.5
Average Amperes (A) 0.6 1.1 2.0 4.3 6.215 V 
STDEV (Amperes) 0.02 0.05 0.05 0.4 0.0
  64  
Table 3b.  Estimated digging energy using fixed voltages of 5, 10 and 15 V.   
 5 V 10 V 15 V 
Estimated Digging 
Energy (per m
3
 of 
sand displaced) 
4.2 MJ 3.2 MJ 2.5 MJ 
Estimated Digging 
Energy (per m
3
 of 
auger buried) 
3.8 MJ 2.9 MJ 2.8 MJ 
 
5.3.2. Sand Mixture  
The results, average time, average current, standard deviation (STDEV) for time 
and current and estimated digging energy are summarized in Table 4a and 4b.  As in 
previous experiments, the digging energy has been calculated based on both the volume 
of sand displaced and depth to which the auger is buried (see 4.2.9 Digging Energy 
Requirement, Equation 1 & 2) 
Table 4a.  Digging results using fixed voltages of 5, 10 and 15 V.  Average values for 
three trials. 
 
 
Begin Digging Auger Quarter 
Buried 
Auger Half 
Buried 
Auger Three
-
Quarters Bu
ried 
Auger Fully 
Buried 
Motor Quarter 
Buried 
Motor Half 
Buried 
Motor Thre
e
-
Quarter Buried 
Final 
Average Time (s) 0 36 98 163 240 340 410 320 500
STDEV (Time) 0.0 9.5 4.5 1.00 1.5 18 43 275 45
Average Amperes (A) 0.5 0.8 1.2 1.7 2.1 2.6 3.1 2.1 3.1
5 V  
STDEV (Amperes) 0.0 0.07 0.03 0.11 0.16 0.23 0.36 1.9 0.43
  
                    
Average Time (s) 0 10 35 60 85 115 135 160 187
STDEV (Time) 0.0 2.1 1.00 1.00 3.5 3.00 3.6 12 12
Average Amperes (A) 0.6 0.8 1.0 1.4 1.8 2.3 2.7 3.1 3.2
10 V 
STDEV (Amperes) 0.02 0.05 0.03 0.03 0.03 0.06 0.19 0.14 0.67
                      
Average Time (s) 0 9 24 40 58 75 86 106 118
STDEV (Time) 0.0 1.5 1.5 1.7 2.1 2.9 4.00 3.8 3.0
Average Amperes (A) 0.6 0.8 1.1 1.6 1.9 2.3 2.6 3.1 2.715 V 
STDEV (Amperes) 0.02 0.06 0.08 0.02 0.09 0.05 0.20 0.13 0.12
  65  
 
Table 4b.  Estimated digging energy using fixed voltages of 5, 10 and 15 V.   
 5 V 10 V 15 V 
Estimated Digging 
Energy (per m
3
 of 
sand displaced) 
1.7 MJ 1.3 MJ .98 MJ 
Estimated Digging 
Energy (per m
3
 of 
auger buried) 
2.8 MJ 2.1 MJ 1.3 MJ 
 
5.3.3. Compacted Sand Mixture 
The results, average time, average current, standard deviation (STDEV) for time 
and current and estimated digging energy are summarized in Table 5a and 5b.  As in 
previous experiments, the digging energy has been calculated based on both the volume 
of sand displaced and the depth to which the auger is buried (see 4.2.9 Digging Energy 
Requirement, Equation 1 & 2) 
Table 5a.  Digging results using fixed voltages of 5, 10 and 15 V.  Average values for 
three trials. 
 
 
Begin Digging Auger Quarter 
Buried 
Auger Half 
Buried 
Auger Three
-
Quarters Bu
ried 
Auger Fully 
Buried 
Motor Quarter 
Buried 
Motor Half 
Buried 
Motor Thre
e
-
Quarter Buried 
Final 
Average Time (s) 0 62 147 232 350 N/A N/A N/A 447
STDEV (Time) 0.0 6.5 19.6 24.1 19.4 N/A N/A N/A 6.5
Average Amperes (A) 0.5 0.8 1.2 1.9 2.5 N/A N/A N/A 3.4
5 V  
STDEV (Amperes) 0.02 0.06 0.06 0.10 0.09 N/A N/A N/A 0.03
  
                    
Average Time (s) 0 29 65 99 133 169 190 N/A 196
STDEV (Time) 0.0 3.0 3.0 3.6 3.6 6.1 6.7 N/A 2.5
Average Amperes (A) 0.5 0.9 1.2 1.9 2.3 2.9 3.2 N/A 3.3
10 V 
STDEV (Amperes) 0.10 0.04 0.07 0.04 0.08 0.22 0.14 N/A 0.06
                      
Average Time (s) 0 16 42 64 88 108 121 130 130
STDEV (Time) 0.0 3.0 7.5 9.5 9.7 11.4 10.8 12.3 12.0
Average Amperes (A) 0.5 0.8 1.2 1.8 2.3 2.7 2.9 3.3 3.015 V 
STDEV (Amperes) 0.04 0.06 0.07 0.08 0.03 0.05 0.16 0.05 0.1
  66  
 
Table 5b.  Estimated digging energy using fixed voltages of 5, 10 and 15 V.   
 5 V 10 V 15 V 
Estimated Digging 
Energy (per m
3
 of 
sand displaced) 
1.7 MJ/m
3
 1.3 MJ/m
3
 1.3 MJ/m3 
Estimated Digging 
Energy (per m
3
 of 
auger buried) 
2.0 MJ/m
3
 1.9 MJ/m
3
 2.0 MJ/m
3
 
 
 
5.3.4. Torque Requirement for Customized Auger Experiments 
In a plain sand bed, with a high friction stress, the custom auger observed a torque 
requirement of 135 Nm, once fully buried.  With the addition of the flour and oil, the 
torque was reduced to only 27 Nm.  However, with a compacted bed consisting of sand, 
flour and oil, the torque increased to 81 Nm.  The results can be found in Table 6. 
 
Table 6.  The torque requirements, measured in Newton-meters, for the custom auger 
testing. 
 Plain Sand Sand Mixture Compacted Sand 
Mixture 
Measured Torque 
(Nm) 
135 (1200 in-lbs) 27 (240 in-lbs) 81 (720 in-lbs) 
 
5.4. Vehicle Experimentation 
5.4.1. Earthworm Concept 
The motor had no problem burying the auger during the initial run, but ceased its 
downward movement once the sand level reached the motor itself.    The voltage was 
kept at a constant 15 V.  The current however, started at 0.6 A but quickly jumped to 
about 2 A, as the sand level reached the motor (Figure 18). Due to the higher torque 
needed, the attachment started to slip around the auger extension instead of rotating the 
  67  
auger.  The conical adapter was then machined to overcome this problem.  Unfortunately, 
due to the limited time and funding available, there was no opportunity to conduct a 
retest. 
 
Figure 18.  Stopping point of the earthworm design. 
 
5.4.2. Friction Reduction 
Tube placement was tested at the nose of the cone attachment, the beginning of the 
motor body and the middle of the motor body.  Placing the tube ends at the nose of the 
cone/top of the auger proved unreliable, as the moving part (adapter) caused interference.  
The rotation of the cone caused the tubes to become horizontal in the sand, thus rendering 
them useless.  Placement in the middle of the body of the motor did not allow full 
involvement of the air.   
  68  
The device was unable to dig far enough down for the tubes to be effective with the air 
pressure and the device stopped well before the tubes were below the surface of the sand.  
These issues led to the decision to keep the tubes level with the beginning of the motor 
(Figure 19).   
 
 
 
Figure 19.  Visual of motor with air hoses to help reduce the surface friction. 
 
 
However, testing at this position was also found to be unsuccessful.  With air blowing 
straight down into the sand, it was thought that there was insufficient air to surface 
interaction to be really effective.  Therefore, small holes were placed at ~ 25 mm, (1 in) 
intervals along the length of the tube, parallel to the surface of the motor.  
  69  
With various pulse intervals, the digger produced the same result; a depth of about 380 
mm (15 in).  This suggests that either the frictional forces were not as big a problem as 
the inability of the auger to dig consistently, or that a more sophisticated air delivery 
system is required.   
5.5. Biomimetic Results 
5.5.1. Weighted Results 
A weighting system was created to assign each item of information a 
corresponding value.  A scale of 1 ? 5 denoted each piece of information as being useful 
to the model (5) or being a hindrance to efficient digging (1).  Seven primary metrics 
were created, with four of the seven based on the total value of the information (speed, 
power, material substance, mechanical complexity) and three of the seven based on the 
appropriateness of the information to a digging scenario (depth, geographic range, size of 
body).  For example, the razor clam can dig in a timely manner for its body size and can 
be given a direct value based on the speed attained.  However, because of its small size 
and nature of its environment (water), these are not appropriate conditions when trying to 
build a large-size digging vehicle for surface applications.  Therefore, it would receive a 
high mark for its speed but low marks for the appropriateness of its size and geographic 
range.  If, however, the scenario were to build a water-based digger, then the razor clam 
would receive high marks for the appropriate geographic range.  As mentioned in section 
1.4.5 Weighting System, each scenario has two scores, with and without arbitrary values.  
The following creatures were identified as worth additional study based on the output 
from the weighting system.  Scenario 1 with arbitrary values produced tied scores 
between the brown rat, narrow faced kangaroo rat and the golden mantled ground 
  70  
squirrel.  The great gerbil was rated the most useful for scenario 2 with arbitrary values.  
The Townsend mole had the best score for scenarios 1 and 2 without arbitrary values.  
The termite is deemed the best candidate for both cases of scenario 3.  Results from all 
the creatures researched can be found in Appendix I.  
It should be noted that none of the information was given a mandatory 
requirement for the scenarios; this is why the termite came out with the highest score for 
scenario 3, deep depths and medium sized body.  The termite was given a low mark for 
its body size however a high mark for its deep depths.  This anomaly is based on the 
amount of information that was found for each creature.  Creatures that fit the 
classification of medium sized body could not dig as deep or did not have enough 
information to warrant a high enough score.  The mole however was of appropriate size 
and is better suited for shallow to medium depths, hence receiving a higher score without 
using arbitrary values.  All the resulting information for each creature can be found in 
Appendix J.
  71  
6. DISCUSSION 
6.1. SMA Testing Procedure 
During testing of the SMAs, many things may have been responsible for 
variations in the experimental procedure that could skew the results.  First, the testing 
was performed with the help of several students; the testing was not performed by the 
same one or two people throughout the life of the project.  This may have led to some 
variation in the readings, for example, how the yard stick was read.  Also, there was 
considerable room for error in the recording of the time at the end point of the 
experiment.  Sufficient time was required to satisfy the observer that all movement of the 
SMA had ceased and no more lifting was going to be accomplished.  While assessing 
this, some observers may have allowed more time than was necessary to complete the 
experiment.  On many occasions, the spring stopped prematurely due to uneven heating 
of the wire.  Once an even distribution of heat had been achieved, the coil would then 
continue to lift the weight.  Because of this phenomenon, it is reasonable to assume that 
different observers would allow different amounts of time to ensure the endpoint had 
been reached.  Another example was the fact that all the springs were tested at the same 
stretched length.  For the smaller springs in the first batch, initial testing at various 
stretched lengths was not performed to determine the optimum working length.  If the 
first batch was overstretched, this could have decreased the SMA?s power. 
The larger springs in the second batch, were tested at the same length as the first batch, 
  72  
although as though they were bigger, they could have theoretically handled a greater 
lifting distance.  By failing to normalize the distance, the second batch could have 
appeared to have a higher efficiency than was actually justified by the results. 
6.2. Auger Testing 
6.2.1. Conical Auger 
The conical auger was designed based on a ?back of the envelope? idea, although 
after viewing the final product, the research team agreed that this was not in fact the best 
approach.  The design produced was un-symmetrical and no preliminary study of the 
literature was conducted to support the shape.  Two augers were fabricated, one using a 
high density polymer and the other bare aluminum.  However, problems were 
encountered as the aluminum was too heavy for the motor to handle.  The polymer auger 
was hard enough to withstand erosion to some extent but it was still affected by the sand, 
leaving many abrasions that were visible to by the naked eye and could be felt by hand.  
The main problem however was the design itself, which required high torques.  This was 
attributed to the wide base and the thick flights spiraling around the center.  These 
produced large surface areas that contributed to the large frictional forces experienced.  
Also, due to its design, the auger would sometimes strip the sand in the borehole, hence 
preventing any further descent.  A visual inspection during the experiments revealed that 
the auger often pushed the sand outward instead of upward and the auger would simply 
spin in place.  Another problem that had to be addressed was that large torques were 
required during the experiments that were too severe for simple anti-counter rotation 
devices to deal with, so a hand held stabilizing mechanism was needed.  The deeper the 
vehicle penetrated, the larger the force needed to stabilize the motor and this became a 
  73  
considerable strain on the operator.  Eventually the strain on the motor would become too 
great, leading to stalling out, in the event that the auger did not strip the borehole. 
6.2.2. Helical Augers 
Industrial helical augers were purchased off the shelf.  The decision to test 
commercially available models to replace the conical auger initially used was dictated by 
their proven track record for post hole digging in real world situations.  However, in 
industrial applications, they are equipped with an ample power supply capable of 
applying sufficient force, weight on bit, and torque, to overcome frictional forces, to push 
the auger down into the earth.  For this experiment, the small diameter (64 mm, 2.5 in 
and 76 mm, 3.0 in) helical augers required a much lower torque to spin through sand 
compared to the conical augers.  This can be attributed to the smaller volume of each 
helical auger.  The helical augers have thin flights and a smaller core diameter which 
greatly reduces the surface area.  However, abrasive markings were still visible evidenced 
by large areas of paint that had been eroded away to expose the base metal.  In contrast, 
the large diameter helical auger (200 mm, 8.0 in) had an extremely high torque value due 
to its larger flights producing a larger surface area and moving a larger amount of sand.  
Both the small diameter helical augers were difficult to use with the available motor and 
a lack of proper attachment also hindered the experiments.  Due to its design, the auger 
tended to strip inside the attachment device used to couple the auger with the motor.  The 
auger?s depth potential were unattainable because of this problem. 
 
 
 
  74  
6.2.3. Custom Auger 
The custom auger finally used for the study was manufactured based on research 
and a design by the University of Missouri Rolla.  This auger was designed to propel the 
vehicle using a self digging mechanism.  The custom auger was the only auger tested 
with the different sand compositions used in the study:  Plain silica sand, a non-
compacted sand mixture and a compacted sand mixture of sand, flour and oil.  Like the 
conical auger first tested, the custom built auger was fabricated out of a high density 
polymer.  However, during the plain sand experiments, this proved to be too big and 
heavy for the motor to drive through the medium.  In the plain silica sand, the vehicle 
was able to bury three quarters of the auger before stalling out.  Again, this stalling out is 
probably due to the large volume and surface area of the auger; the motor was simply not 
powerful enough to overcome the frictional forces encountered.  In the non-compacted 
sand mixture experiments, the vehicle was able to fully bury the auger and three quarters 
of the motor.  However, these results may not be a reflection of the true ability of the 
vehicle.  Due to the low compaction of the mixture, the vehicle would sink in a few 
inches due to its own weight when simply placed on top of the medium.  The mixture of 
the flour, sand and oil was very porous and thus gave the vehicle an extra advantage.  
This led to the experiments being repeated in the same sand mixture that had been highly 
compacted and was less porous.  The results of the new runs showed that the vehicle was 
unable to dig as far as previous non-compacted experiments at low speeds, although it 
was able to reach previous depths when a full 15 V was supplied to the motor.  With 
regard to the amount of torque needed to make the auger spin, the plain sand experiments 
were observed to require the highest torque followed by the compacted mixture and the 
  75  
non-compacted mixture, falling in that order.  The plain sand resulted in the highest 
friction due to the dry silica particles rubbing against the surface of the auger, while the 
addition of flour and oil helped reduce the friction between the medium and the auger 
surface.   
6.3. Vehicle Experiments 
6.3.1. Earthworm Concept 
This experiment was run before the adapter shown in Figure 6 became available 
and therefore lacked a smooth transition surface between the auger and motor.  The 
discontinuity, caused by the flat face of the motor, resulted in the sand becoming 
compacted at the point of transition from auger to motor, at which time the digging action 
ceased.  It is likely that the resulting slippage of the attachment between the motor and 
the auger was the main reason for the subsequent lack of downward progress. 
6.3.2. Air hose/friction reduction 
A range of tube placements were tested:  At the nose of the cone attachment, the 
beginning of the motor body and the middle of the motor body.  At the nose of the cone 
attachment, there was no good way to fix the tubes in place that would allow them to 
remain stationary, so the tubes were forced to rotate with the cone if fixed to the surface.  
As a result, they flattened outward and curved upward as the vehicle penetrated into the 
sand, rendering them useless.  When mounted in the middle of the motor body, the 
experiments generated unrepeatable results and in some runs even failed to penetrate far 
enough for the air to play an active role.  In the experiments in which the vehicle did dig 
deep enough, the effect of the air hose was minimal.  In the final location tested, at the 
beginning of the motor, it was observed that the air was focused into small areas at the 
  76  
ends of the hoses and was not able to provide any assistance to the other areas of the sand 
to motor surface interface.  In a further attempt to reduce the friction at the motor surface, 
small holes were drilled along the length of the air tubes.  As the motor is of a radial 
design, the holes were not parallel with the surface, but at a tangent.  A large number of 
design issues remained to be overcome in order for this system to be effective, so as with 
the experiments with fin design, this was deemed a secondary issue that could be returned 
to in future studies. 
6.4. Biomimetic System 
As mentioned in the previous chapter, various rodents were selected using the 
weighting system as warranting further examination for a study of the mechanisms they 
use to dig at shallow depths with arbitrary values.  However, these rodents are not 
suitable as a model for this study because they dig with their claws and hind feet.  They 
have no specifically adapted body parts, but are simply able to dig due to their small size 
and high strength.  An investigation into replicating the movements and body of the mole 
may be more productive.  The body of the mole consists of a tapered head, a streamlined 
body and highly adapted arms, all of which contribute to the mole?s ability to dig quickly 
and powerfully. 
 The mole?s arm and paw are highly adapted for digging.  Its humerous bone is not 
like the long thin bones typically found in other mammals, but is as wide as it is long, 
making it almost rectangular in shape.  To be able to handle moving large amounts of 
soil, the mole?s back muscles (trapezious and latissimus dorsi) are directly connected to 
the humerous bone.  In humans, these muscles indirectly affect the ability to move and 
lift with the arms.  This muscle structure is unlike any other, and allows the paw to 
  77  
extend out to the sides of the body and push directly back. The paw is already oriented 
outward with thick, long claws that easily penetrate soil.  By pushing the soil outward, 
moles overcome the problems of sand movement.  Other animals have to first loosen the 
soil, and then push it underneath their bodies and out of the tunnel.  This is inefficient, as 
many strokes are needed.  Moles can push their body into the material and then move the 
material behind them in one smooth stroke using a movement akin to swimming.  After 
extending their arm and pushing the soil out to the sides, moles are then able to draw their 
paws close to the body and out in front of their face in tight spaces.  This is only possible 
due to the unique design of the humerous bone. 
 The body is also of a desirable size as it is large enough so that scaling up results 
in no loss of the power to weight ratio, unlike in insects.  One attribute that is highly 
useful is the mole?s ability to turn 180 degrees inside relatively small tunnels.  The 
tunnels moles construct are only a little more than the width of their body, although their 
body length is much longer.  It is unknown how moles are able to contort their bodies to 
be able to accomplish this feat.  The fur on the outside of the mole?s body also helps in 
their movements.  It is non-oriented, and this lack of a specific direction allows it to move 
in all directions freely, thus reducing the drag between the soil particles and the mole?s 
body.  
 
 
  78  
7. CONCLUSIONS 
7.1. Shape Memory Alloys 
The use of SMA to provide the mechanical power for a digging vehicle is a novel 
approach, but the design of the springs that are currently available leaves a lot to be 
desired.  Large strokes can be obtained at low or no loads but were inefficient. At greater 
loads more energy input is necessary and considerably smaller strokes were achieved.  
Testing of each spring?s power by recording the current and the subsequent time needed 
and distance traveled to lift a given amount of weight revealed that the efficiency is 
unsatisfactory for repeatable use and the stroke length decreased with every experiment.  
The fabrication of larger springs capable of lifting large loads increased the efficiency, 
but the ?memory loss? effect became even more of an issue. 
The first batch of springs, which were the smaller of the two sets, produced 
efficiencies (of electrical power to mechanical work transformation) as low as 0.003% 
with a maximum of 0.19%.  The second batch of springs, which were larger, produced 
efficiencies between 0.23 ? 0.56%.  Although this is an improvement over the smaller 
springs, it is still well below the average energy conversion efficiency for a ?typical? 
NiTi SMA, at 5%.  One possible explanation for this low efficiency was that the heating 
cycled employed for the work was not optimized for the wires used. It is possible that the 
use of higher currents, for example 9-10 A, would decrease the amount of time needed 
for the wire to go through its transformation.   
  79  
Further study is needed to determine the optimum balance between the two variables of 
current and applied force.  Very little loss of the material?s memory was observed in the 
first batch and the results were very repeatable.  However the current and force applied 
for this phase were much lower, at a maximum of 6 A and 22 N, respectively, compared 
to those of the larger springs where testing started at 10 A and 45 N, respectively.  The 
second batch showed signs of memory loss while testing was in progress and the 
effectiveness of the strokes decreased markedly with each successive run.  These 
observations confirm those seen in previous experiments; although, the magnitude of the 
decline was not as extreme as those seen by previous recordings.  Thus, it is possible that 
the lower currents were still deleterious but took longer to cause extreme damage.  These 
two reasons make SMA springs an unpromising candidate to power a reusable digging 
vehicle.   
7.2. Augers 
7.2.1. Conical Augers 
In measuring the amount of torque needed to drive the 127 & 140 mm conical 
augers through sand, they required the second highest values, exceeded only by the 203 
mm diameter helical auger.  During the colored sand experiment, it was observed that the 
sand was not displaced upward by the auger?s action but outward.  The thick flights, 
conical shape and overall design proved to be unreliable when used in a sand medium. 
7.2.2. Helical Augers 
The two similar diameter helical augers (63 mm, 76 mm) produced the smallest 
torque values, and succeeded in moving through the sand medium with less than 20 Nm 
to a depth of only 0.6 meters.  The largest of the helical augers (203 mm diameter) 
  80  
provided an excellent demonstration of how much the required torque increases as size 
increases, with over 160 Nm needed to penetrate half the depth (0.3 m) achieved by the 
two smaller augers.  Unfortunately the amount of torque needed to move the large 
diameter helical auger exceeded what was possible using the motor that was available for 
the tests.  The smaller diameter augers proved more suitable for this medium, although 
greater depths were not achievable as the vehicle?s descent ceased due to a lack of weight 
on the bit or due to design issues such as the motor to auger attachment.  Without 
additional weight, the vehicle experienced a type of buoyancy as the weight and power of 
the vehicle balanced the frictional forces needed to keep moving.  If additional force was 
applied from above the vehicle would continue to dig, but once the force was removed 
the auger would again spin in place.  A suitable attachment was also difficult to achieve 
with the helical augers.  The thin center core made keeping the auger stationary with 
respect to the motor complicated, and as a result the augers tended to strip the inside of 
the attachment devices used to couple the auger with the motor.  The design of the helical 
augers was therefore deemed unsuitable for a self propelled digging vehicle.   
7.2.3. Custom Auger 
The custom auger tested in this study was designed specifically for digging in 
sand and was tested using three different sand media:  plain silica sand and a mixture of 
sand, flour and oil that was either compacted or un-compacted.  All the tests used fixed 
voltages of 5, 10 and 15 V.  The plain sand trials proved to require the most power as the 
estimated digging energy was higher than that for the other two trials for all three of the 
fixed voltages tested.  The non-compacted sand mixture did not constitute a realistic 
experiment as the porous medium allowed the vehicle to start to sink, thus artificially 
  81  
increasing the depth achieved, due to its weight alone, before digging even started.  The 
torque in the plain sand experiments rivaled that of the conical auger at 135 Nm, but was 
significantly less in both the sand mixtures.  The custom auger thus presented a viable 
alternative to the other augers tested, but as the focus of the project changed, the auger 
tests did not proceed further. 
7.3. Biomimetic System 
The time frame of the project allowed a detailed and descriptive review of a range 
of biological organisms to be conducted.  In order to maximize the data gathered without 
focusing on the lack of data for any one organism, a small database of fossorial creatures 
was established.  Using the available knowledge for each organism, a weighted system 
was created that allowed the output of different animals for different scenarios to be 
compared.   
Even though the mole did not out perform the other animals for all the theoretical 
environments considered, its muscular and frame structure deserve a closer look as part 
of an attempt to construct a new model for a digging vehicle.  The weighting system  
used in this comparison was based on speed, depth, power, material substance, 
mechanical complexity, geographic range and the size of the body of the creature, and 
confirmed that the mole was the best candidate out of the 43 examined for depths of less 
than 6 m (240 in) and a body length of less than .3 m (12 in).  The mole?s body size, 
geographic range and digging depth are well suited to vehicle modeling.  The mole has 
many attributes that can be viewed as highly desirable, such as its unique digging 
movement, the large aspect ratio of its digging apparatus, and its cylindrical and 
streamlined body structure.   
  82  
The researcher therefore recommends that future research should focus on the Townsend 
mole, rather than any of the other organisms surveyed. The results for the mole can be 
found in Appendix K.   
  83  
8. FUTURE RESEARCH 
Many of the augers tested in this study were not designed specifically for self 
propulsion through a sand media.  The UM-Rolla (UMR) auger, however, is designed to 
do so and has the potential to dig to great depths.  Although the UMR auger discussed in 
this thesis was much too large for the available motor to handle, a scaled down version 
might accomplish the research objective, namely constant and steady penetration.  By 
scaling down the size of the auger, the motor will need less power to turn it and therefore 
will have more in reserve to compensate for the increasing frictional forces at greater 
depths.  The vehicle?s digging ability could also be enhanced through the addition of a 
vibration mechanism.  The use of a fluid media might also prove beneficial.  The use of 
air to boost the auger?s digging action in this study was a quick and rough experiment 
that returned only an inconclusive result.  This approach may merit a more detailed and 
better researched trial.  The use of vibration and/or fluid media in various other digging 
devices are known to assist in the digging process by removing the digging media, such 
as the use of water based fluids in oil drilling platforms.   
The next step in the effort to advance the research into biomimetic devices is to 
determine the exact mechanism by which a mole?s swimming motion propels it through 
the earth and measure the forces involved.  This would enable detailed models to be 
constructed using various body sizes and speeds.   
  84  
This thesis has provided only the very beginning stages of researching new and 
innovative approaches to mechanical digging; much more research, design and testing is 
needed to accomplish a comprehensive theoretical treatment and working demonstration 
model.  
  85  
REFERENCES 
 
1. Maciejewski, J. and Jarzebowski, A.; ?Laboratory optimization of the soil digging 
process,? Journal of Terramechanics, vol. 39, 2002, pp. 161-179. 
2. Sellmann, P.V. and Brockett, B.E.; ?Bit Design Improves Augers,? The Military 
Engineer, Vol. 79, August, 1987, pp. 453-454. 
3. Chugh, C.P.; Manual of Drilling Technology, A.A. Balkema/Rotterdam, India. 
1985, pp. 1-45. 
4. Onshore Drilling, http://www.naturalgas.org/naturalgas/extraction_onshore.asp, 
2005 
5. Petroleum Extraction, http://www.personal.psu.edu/users/c/b/cbm134/history.htm, 
2005 
6. Institute for Energy Research, 
http://instituteforenergyresearch.org/Energy_Timeline.htm, 2005 
7. Maciejewski, J., Jarzebowski, A. and Trampczynski, W.; ?Study on the efficiency 
of the digging process using the model of excavator bucket,? Journal of 
Terramechanics, vol. 40, 2004, pp.221-233. 
8. Abo-Elnor, M., Hamilton, R. and Boyle, J.T.; ?Simulation of soil-blade 
interaction for sandy soil using advanced 3D finite element analysis,? Soil & 
Tillage Research, vol. 75, 2004, pp. 61-73. 
9. Georgiadis, K., Potts, D.M. and Zdravkovic, L.; ?Modelling the shear strength of 
soils in the general stress space,? Computers and Geotechnics, vol. 31, 2004, pp. 
357-364. 
10. Finite Element Analysis Modeling Issues and Ideas, 
http://www3.sympatico.ca/peter_budgell/Modeling_issues.html#C1, 2005 
11. Anonymous; ?Trenchless technology speeds gas line renewals?, Pipeline & Gas 
Journal, vol. 222, 1995, pp.23-25.
  86  
12. Stidger, Ruth. W; ?Technology:  Trenchless technology provides environmental 
advantages?, Gas Utility Manager, vol. 47, November, 2002. p.18-19. 
13. Friant, J.E. and Oxdemir, L.; ?Tunnel Boring Technology ? Present and Future?, 
Proceedings of the Rapid Excavation and Tunneling Conference, 1993, pp.869-
888. 
14. Worthington, T.; ?New Tools for Tunneling, Finding Niches in Hardrock?, 
Engineering and Mining Journal, vol. 204, February, 2003, pp.32-34. 
15. Normile, D.; ?Cutting Corners,? Popular Science, vol. 246, April, 1995, pp 27. 
16. Hill, J.L. III, Shenhar, J. and Lombardo, M.; ?Tethered, Down-Hole-Motor 
Drilling system ? A benefit to Mars exploration,? Advanced Space Research, vol. 
31, 2003, pp.2421-2426  
17. Krajick, K.; ?New drills augur a great leap downward? Science, vol. 283, 1999, 
pp.781.   
18. Mavroidis, C.; Pfeiffer, C. and Mosley, M.; ?Conventional actuators, shape 
memory alloys and electrorheological fluids?, Research in Nondestructive 
Testing, vol. 14, 2002, pp.1-32. 
19. Shape Memory Alloys,  
http://www.cs.ualberta.ca/~database/MEMS/sma_mems/sma.html, 2005 
20. Gale, W.F.; Auburn University Materials Research and Education Center, 
personal communication. 
21. Introduction to Shape Change and Superelasticity,   
http://www.jmmedical.com/html/introduction.html, 2005. 
22. Huang, W.; ?On the selection of shape memory alloys for actuators?, Materials 
and Design 23, 2002, pp.11-19. 
23. Park, B., Shantz, M. and Prinz, F.; ?Scalable rotary actuators with embedded 
shape memory alloys?, Proc. Of The International Society for Optical 
Engineering, vol. 4327, 2001, pp.79-87. 
24. Paulson, L.; ?Biomimetic robots?, Computer Magazine, September 2004, pp.48-
53. 
25. http://www.first-to-fly.com/History/History%20of%20Airplane/history.htm, 2005 
  87  
26. http://inventors.about.com/library/weekly/aa091297.htm, 2005 
27. Vanderbuilt, T.; ?Evolution of the gearhead?, DigitAll Magazine, summer 2004, 
http://www.samsung.com/DigitAll/BrandCampaign/magazinedigitall/2004_summ
er/feat_04a.htm 
28. Isla, D.; ?RoboTuna project to model real fish?, The Tech Magazine, Vol. 115, 
Number 49, 1995, pp. 13, 23.   
http://www-tech.mit.edu/Issue/V115/N49/robotuna.49n.html 
29. Sheenan, C.; ?Robotics gains in prestige, in part due to military conflicts? 
Associated Press April 2004.  
http://www.aaai.org/AITopics/html/archvE4.html#april8b 
30. Brain, M.; Personal BLOG: Robotic Nation Evidence, article ?Formerly arcane 
research gets new respect? April 2004 
http://roboticnation.blogspot.com/2004/04/future-of-robotic-police.html 
31. Weed, W.; ?This is your computer speaking?? Discover Magazine, Volume 23 
No. 8 August 2002. 
http://www.aaai.org/AITopics/html/autveh.html 
32. News office; ?MITs robotic helicopter makes first aerobatic roll? February 2002 
http://web.mit.edu/newsoffice/2002/robochopper.html 
33. DARPA Grand Challenge 05  
http://www.darpa.mil/grandchallenge/overview.html 
34. http://www.vattenkikaren.gu.se/fakta/arter/mollusca/bivalvia/ensiamer/ensiam2e.h
tml, 2005 
35. http://www.ecy.wa.gov/programs/sea/pugetsound/species/geoduck.html, 2005 
36. http://www.liveaquaria.com/product/prod_Display.cfm?siteid=23&pCatId=579, 
2005 
37. http://www.chesapeakebay.net/shipworm.htm, 2005 
38. http://molecrickets.ifas.ufl.edu/mcri000b.htm, 2005 
39. http://molecrickets.ifas.ufl.edu/mcri0001.htm, 2005 
40. http://www.dpi.qld.gov.au/beef/13129.html, 2005 
41. http://www.ces.ncsu.edu/depts/ent/clinic/Bugofwk/970081/dungbeet.htm, 2005 
  88  
42. http://ianrpubs.unl.edu/Insects/g1062.htm, 2005 
43. http://www.scienceinafrica.co.za/2004/april/trapdoor.htm, 2005 
44. http://www.zo.utexas.edu/research/txherps/frogs/rhinophrynus.dorsalis.html, 2005 
45. http://animaldiversity.ummz.umich.edu/site/accounts/information/Rhinophrynidae
.html, 2005 
46. http://animaldiversity.ummz.umich.edu/site/accounts/information/Perognathus_fl
avescens.html, 2005 
47. http://animaldiversity.ummz.umich.edu/site/accounts/information/Perognathus_fa
sciatus.html, 2005 
48. http://animaldiversity.ummz.umich.edu/site/accounts/information/Thomomys_tal
poides.html, 2005 
49. http://animaldiversity.ummz.umich.edu/site/accounts/information/Geomyidae.htm
l, 2005 
50. http://animaldiversity.ummz.umich.edu/site/accounts/information/Heterocephalus
_glaber.html, 2005 
51. http://animaldiversity.ummz.umich.edu/site/accounts/information/Scapanus_town
sendii.html, 2005 
52. http://sandia.gov/media/NewsRel/NR2001/minirobot.htm, 2005 
53. Mondada, F. and Guignard, A.; ?SWARM-BOT: From Concept to 
Implementation?, Proceeding of the 2003 IEEE/RSJ.  Intl. Conference on 
Intelligent Robots and Systems, Las Vegas, Nevada, October 2003.   
54. Kahn, J.M., Katz, R.H. and Pister, K.S.J.; ?Next Century Challenges:  Mobile 
Networking for ?Smart Dust?; Proceeding of the 1999 5
th
 Annual ACM/IEEE 
International Conference on Mobile Computing and Networking (MobiCom?99),  
1999, p.271-278. 
55. Stormont, D.P.; ?Robot swarms for planetary exploration?, Proc. 4
th
 International 
Conf. and Exposition/Demonstration on Robotics for Challenging Situation and 
Environments, 2000, p.347-352. 
  89  
56. Fukuda, T., Mizoguchi, H., Sekiyama, K. and Arai, F.; ?Group behavior control 
for MARS (Micro Autonomous Robotic System)?, Proc. 1999 IEEE International 
Conf. on Robotics and Automation, May 1999.  pp.1550-1555. 
57. Kadrovach, B.A., and Lamont, G.B.; ?A particle swarm model for swarm-based 
networked sensor systems?, Proc. Of the ACM Symposium on Applied 
Computing, 2002 p.918-924. 
58. Kube, C.R. and Bonabeau, E.;  ?Cooperative transport by ants and robots?, 
Robotics and Autonomous Systems, vol. 30, 2000, p.85-101 
59. Fukuda, T., Funato, D., Sekiyama, K. and Arai, F.; ?Evaluation on flexibility of 
swarm intelligent system?, Proc. 1998 IEEE International Conf. on Robotics and 
Automation, May 1998, p.3210-3215. 
60. Dudek, G., Jenkin, M.; ?A taxonomy for swarm robots?, Proc. 1993 IEEE/RSJ 
International Conf. on Intelligent Robots and Systems, July 1993, p.441-447. 
61. http://www.sandia.gov/media/NewsRel/NR2000/avalanch.htm. 
62. Russell, R.A.; ?Pheromone communication: Implementation of necrophoric bee 
behavior in a robot swarm?, Proc. 2004 IEEE Conf. on Robotics, Automation and 
Mechatronics, December 2004, p.638-643. 
63. Lam, Y.K., Wong, E.K. and Loo, C.K.; ?Explicit communication in designing 
efficient cooperative mobile robotic system?, Proc. 2003 IEEE International Conf. 
on Robotics and Automation, September, 2003. p.3869-3874. 
64. Owusu, Yaw A.; ?Physical-chemical study of sodium silicate as a foundry sand 
binder?, Advances in Colloidal Interface Science, September, 1982. p.57-91. 
65. Cash, Alex ; Air Force Research Lab, Eglin Air Force Base, Fl., personal 
communication. 
 
 
 
 
 
 
  90  
APPENDICES
  91  
Appendix A.  Draft CAD Drawing of Conical Auger 
 
Figure A-1:  View of Conical Auger  
  92  
Appendix A.  Draft CAD Drawing of Conical Auger 
 
 
Figure A-2:  View of Conical Auger 
  93  
Appendix B.  Auger and Motor with Fin Design 
 
 
Figure B-1:  Motor with three small fins 
 
 
Figure B-2:  Motor and conical auger with two large fins 
  94  
Appendix B.  Auger and Motor with Fin Design 
 
 
Figure B-3:  Motor and conical auger with rod-fin design 
  95  
Appendix C.  Extended Auger 
 
Figure C-1.  Extended auger used for pull test 
 
  96  
Appendix D.  Draft CAD Drawing of Custom Auger 
 
Figure D-1.  Side view of custom auger, designed by Dr. Baird from UM-Rolla. 
  97  
Appendix E.  List of Initial Questions 
List of questions pertaining to insects/mammals/fish/amphibians that ?dig? 
? Does the organism carry the material displaced away? 
? Does the organism gain sustenance from the material? 
? If the organism digs mechanically, then: 
o Does the organism push or scrape at the dirt? 
? If scrape, then: 
? Brief description of digging apparatus (e.g. claws) 
? Any interesting characteristics of the digging 
apparatus? 
? Aspect ratio of the digging apparatus? 
? Does the same apparatus serve other functions? 
o Defense 
o Utility 
? If push, then: 
? Does the organism push the material to the side? 
? Does the organism push the material behind its body? 
? Other movements? 
? Which limbs does the organism use? 
o Front, back or both? 
? Does the organism use any other body parts to  assist 
with digging? 
o How far does the organism dig? 
? Just enough to cover itself? 
? Further? 
? If so, estimated depth by body length. (e.g.  2 times its 
length) 
? Estimated speed (absolute or relative to body length)? 
o Why does the organism dig? 
? Protection 
? Food 
? Hibernation 
? Housing/shelter 
? Other 
o Does the organism need/use help from symbiotic organisms? 
? If so, how do they help each other? 
? How closely related are the species involved? 
? How big are its limbs compared to its body? 
o Describe how the limbs move 
? Swimming motion 
? Overhand motion 
? Multi-jointed 
 
 
  98  
Appendix E.  List of Initial Questions 
 
? How much weight can the organism carry/push/move compared to its own 
body weight? 
o Example: an ant can carry ~10X its body weight 
? Estimated size of organism (dimensions) 
 
Questions pertaining to basic organism information 
? What unique capability/function helps the insect survive? 
? What type of environment does the organism live in? 
o Can the organism survive in a low nutrient (e.g. almost pure water) 
environment? 
o Nocturnal?  
? Average life span? 
? Major sensing mechanism? 
o Eyes 
o Antennae 
o Vibration 
o Other 
? How does the organism communicate? 
o Chemical 
o Vocal 
o Movement/gesture 
? Defense Mechanisms? 
o Stealth/protective coloring 
o Odor 
o Chemical Secretion 
o Claws etc. 
o None 
o Other 
? Eating habits? 
o Carnivore 
o Herbivore 
o Omnivore 
o Other (e.g. gains energy from soil or converts sunlight as in 
photosynthesis) 
  99  
Appendix F.  Scenarios for Weighted Values 
 
Table F-1:  Weighted values for three scenarios 
 
Weightings 
 
Scenario 1 (shallow 
depths, small body 
size) 
Scenario 2 (medium 
depths, small body 
size) 
Scenario 3 
(deep depths, 
medium body 
size) 
Size of body  30% 25% 15%
Speed 25% 15% 10%
Mechanical Complexity 15% 15% 10%
Hard Substances 10% 10% 10%
Geographic Range 10% 10% 5%
Power 5% 10% 30%
Depth  5% 15% 20%
Totals 100% 100% 100%
Weighted = Arbitrary values included    
Weighted = Only known values     
Data Input    
Maximum value output for scenario    
Scenario 1 
Size of Body (<30cm; 12in)    
Depth (<152cm; 60in)    
    
Scenario 2 
Size of Body (<30cm; 12in)    
Depth (152cm;60in<X<610cm;240in)    
 
Scenario 3 
Size of Body 
(30cm;12in<X<91cm;36in)    
Depth (>610cm;240in)
 
  100  
Appendix G.  Biometric Flowchart 
 
Figure G-1:  Biomimetic Flowchart 
  101  
Appendix H.  SMA Spring Data 
 
Table H-1.  Results of Spring 1 with 44.5 N at various currents. 
   
44.5N (10 lbs) 
Observation Time (s) Distance (mm)
  20 A 
1 12 170 
2 15 165 
3 15 160 
  18 A 
1 14 155 
2 12 150 
3 12 150 
  16 A 
1 12 150 
2 14 140 
3 14 140 
  14 A 
1 17 135 
2 16 135 
3 15 130 
  12 A 
1 23 130 
2 21 125 
3 24 130 
  10 A 
1 36 125 
2 35 125 
3 35 125 
 
 
 
  102  
Appendix H.  SMA Spring Data 
Table H-2.  Results of Spring 2 showing the differences in actuation distance (mm) with 
a load of 44.5 N at various currents 
   
44.5N (10 lbs) 
Observation Initial Test 2nd Testing
  10 A 
1 150 105 
2 140 105 
3 140 105 
  12 A 
1 140 110 
2 140 110 
3 135 110 
  14 A 
1 140 115 
2 140 115 
3 140 115 
  16 A 
1 135 115 
2 135 115 
3 135 115 
 
. 
Table H-3.  Results of Spring 3 showing the differences in actuation distance (mm) at 
low currents with a load of 26.7 N 
   
26.7N (6 lbs) 
Observation Initial Test 2nd Testing
  9A 
1 195 170 
2 190 170 
3 185 170 
  8A 
1 175 165 
2 175 165 
3 170 165 
  7A 
1 170 160 
2 165 160 
3 165 155 
 
. 
  103  
Appendix I.  Final Results for all Creatures 
Table I-1.  Final Results for all Creatures 
Specimens Scenario 1: Scenario 1: Scenario 2: Scenario 2: Scenario 3: Scenario 3: 
Razor Clam 6.9 3.3 6.9 4.2 7.9 3.7 
Geoduck Clam 6.5 3.65 5.95 3.7 6.65 2.75 
Common Piddock 7.95 4.05 7.7 4.1 9.3 3.9 
Peanut Worm 6.45 2.55 6.05 2.45 7.3 1.9 
Ship Worm 7.05 3.45 6.75 3.45 8.5 3.1 
June Bug/Grub 6.65 3.05 5.85 3.15 6.7 2.5 
Dung Beetle 8 4.25 6.8 3.95 7.5 3.3 
Mole Cricket 7.05 3.45 6.75 4.05 7.7 3.5 
Termite 8.5 5.2 7.9 5.5 9.9 5.7 
Cicadas 6.75 3.15 6 3.3 6.9 2.7 
Trapdoor Spider 6.75 3.6 6 3.75 7.2 3 
Carpenter Bee 6.45 3.9 5.95 4 7.55 3.65 
Blind Snakes 6.85 3.25 6.3 3.6 7.1 2.9 
Mexican Burrowing Toad 6.85 2.2 6.45 2.1 7.3 1.6 
Tiger Salamander 7.95 4.2 7 4.15 7.35 2.85 
Plains Pocket Mouse 7.75 4 6.65 4.1 7.45 2.95 
Olive Backed Pocket Mouse 7.95 4.2 7.25 4.7 8.05 3.55 
Brown Rat 8.55 4.5 7.75 4.3 8.55 2.85 
Narrow-Faced Kangaroo Rat 8.55 4.5 7.75 4.3 8.55 2.85 
Gulf Coast Kangaroo Rat 7.75 4.45 7.2 4.2 8.2 2.8 
Malagasy Giant Rat 8.45 4.7 7.9 5.35 8.55 4.05 
Yellow Cheeked Vole 7.3 4.3 6.35 4.25 7.25 3.05 
Long Tailed Vole 7.95 3.9 7.25 3.8 8.25 2.55 
Conovers Tuco Tuco 7.6 4.6 6.6 4.5 7.55 3.35 
Golden Mantled Ground Squirrel 8.55 4.8 7.6 5.05 8.15 3.65 
Washington Ground Squirrel 7.75 4.45 7.2 4.2 8.2 2.8 
Idaho Ground Squirrel 7.7 4.7 6.9 4.8 7.6 3.4 
Northern Pocket Gopher 7.75 4.45 7.2 4.8 7.45 3.25 
Western Pocket Gopher 7.75 3.7 6.65 3.8 7.25 2.75 
Valley Pocket Gopher 8.25 4.2 7.65 4.8 8 3.5 
Camas Pocket Gopher 8.05 4.45 7.45 4.15 8.15 2.75 
Black Capped Marmot 7.65 2.7 7.2 2.55 8.45 2.45 
Fat Tailed Gerbil 6.8 3.65 6.15 3.9 7.2 3 
Great Gerbil 8.35 4.75 8 5.3 8.75 4.55 
Indian Gerbil 7.3 4.15 6.35 4.1 7.15 2.95 
Townsend Mole 7.95 5.7 7.5 5.85 7.95 4.05 
Hairy Tailed Mole 7.6 5.35 6.65 5 7.35 3.45 
Pink Fairy Armadillo 7.45 4.15 6.95 3.95 8.05 2.65 
Giant Armadillo 7.15 3.85 6.7 3.7 8.5 3.1 
African Striped Weasel 8.05 4.75 7.45 4.45 8.2 2.8 
Stink Badger 7.75 4.45 7.2 4.2 8.35 2.95 
Naked Mole Rats 8.25 5.1 7.6 5.35 8.2 4 
African Mole Rat 8.45 5 7.55 4.4 8.2 2.8 
  104  
Appendix J.  Information Found for Each Creature 
 
Creature Specific Information Found 
  
Digging 
depths (cm) 
Depth to body 
length ratio 
Body 
Length 
(cm) 
Digging 
Material Speed 
Razor Clam 150-600 20-40 8-15 loose soil   
Geoduck Clam 91 5 19 water,mud 
30 
cm/yr 
Common Piddock     12 rock   
Peanut Worm     1-7 
water, 
loose soil   
Ship Worm     60 wood   
June Bug/Grub 10 8 1 top soil   
Dung Beetle 30 6 1-5 
compact 
soil   
Mole Cricket 152 30 4-5 
loose soil, 
sand   
Termite 4570 >25 <1 wood, soil   
Cicadas 30-60 12 2-5 
compact 
soil   
Trapdoor Spider 30 12 2-5 loose soil   
Carpenter Bee 8 20 2-5 wood 
2 cm/6 
days 
Blind Snakes 200 10 15-20 loose soil   
Mexican Burrowing 
Toad     5-7 loose soil   
Tiger Salamander 60 2 17-33 loose soil   
Plains Pocket Mouse 15-20 5 14 loose soil   
Olive Backed Pocket 
Mouse 200 14 12-14 loose soil   
Brown Rat     32 loose soil   
Narrow-Faced 
Kangaroo Rat     32 loose soil   
Gulf Coast Kangaroo 
Rat     20-27 loose soil   
Malagasy Giant Rat 500 14 30-35 loose soil   
Yellow Cheeked Vole 30 2 14-18 loose soil   
Long Tailed Vole     10-18 loose soil   
Conovers Tuco Tuco 30 1 33-44 loose soil   
Golden Mantled 
Ground Squirrel 100-200 6 30 loose soil   
Washington Ground 
Squirrel     19-25 loose soil   
Idaho Ground 
Squirrel 50-121 5 21-26 loose soil   
Northern Pocket 
Gopher 200 10 21 loose soil   
Western Pocket 
Gopher 10-15 1 15 loose soil   
  105  
Appendix J.  Information Found for Each Creature 
 
 
Digging depths 
(cm) 
Depth to body 
length ratio 
Body Length 
(cm) 
Digging 
Material Speed 
 
Valley Pocket 
Gopher 
 
100-300 
 
10 
 
11-33 
 
loose soil   
Camas Pocket 
Gopher     33 loose soil   
Black Capped 
Marmot     53-73 
compact 
soil   
Fat Tailed Gerbil 100 9 11 loose soil   
Great Gerbil 150-250 15 15-20 
clay, 
compact 
soil   
Indian Gerbil 50 3 15-17 loose soil   
Townsend Mole 300 14 21 loose soil 
5.5 
m/hr 
Hairy Tailed 
Mole 45-70 5 12-14 loose soil 
4.5 
m/hr 
Pink Fairy 
Armadillo     13-15 loose soil   
Giant Armadillo     90 
compact 
soil   
African Striped 
Weasel     25-35 loose soil   
Stink Badger     39 loose soil   
Naked Mole Rats 200 20 10 loose soil   
African Mole Rat     11-17 loose soil   
 
 
  106  
Appendix K.  Metric Results for the Townsand Mole 
 
Table K-1.  Metric Results for the Townsand Mole 
Speed     
Length to width ratio of digging apparatus  
Ratios   Value Known Values 
0.9-1.9  5  
0.89-0.8 2.0-2.9  4 
0.79-0.7 3.0-3.9 3  
0.69-0.6 4.0-4.9  2  
0.59-0.5 5.0-5.9 1 
<0.5 >6  0  
  Given Value: 3 0
Size of digging apparatus to size of body ratio  
Ratios   Value Known Values 
25-30%  5  
20-24% 30-34%  4 
15-19% 35-39%  3  
10-14% 40-44%  2  
5-9% 45-49%  1 
<5% >50% 0  
  Given Value: 3 0
Number of strokes per unit time  
Strokes per minute  Value Known Values 
>25  5  
20-25   4 
15-20  3  
10-15   2  
1-10  1 
  Given Value: 3 0
Coordinated vs random team effort  
Type     Value Known Values 
Coordinated effort  5  
Unknown   3  
Random effort  1 solitary 
  Given Value: 1 1
Raw Speed  
Rate   Value Known Values 
Very Fast   5  
Slightly Fast  4 5.5m/hr 
Intermediate  3 
Slightly Slow  2  
Very Slow   1  
  Given Value: 4 4
  107  
Appendix K.  Metric Results for the Townsand Mole 
 
Distance moved in one stroke  
Amount   Value Known Values 
>50% body length  5  
35-49% body length  4  
20-34% body length  3  
5-19% body length  2  
1-4% body length  1  
  Given Value: 3 0
     
17 5
 
Depth       
Applied force  
Force   Value Known Values     
>85% of its body weight 5      
70-84% of its body weight 4      
55-69% of its body weight 3  
40-54% of its body weight 2      
25-39% of its body weight 1      
<24% of its body weight 0      
  Given Value: 3 0   
Depth to body length ratio    
Overall ratio  Value Known Values     
>25X body length  5    
20-25X body length  4      
15-20X body length  3      
10-15X body length  2 14x  
1-10X body length  1      
  Given Value: 2 2   
  108  
Appendix K.  Metric Results for the Townsand Mole 
 
Appropriate depth Scenario 1 Scenario 2 Scenario 3 
Appropriate level  Value 
Known 
Values Value 
Known 
Values Value 
Known 
Values 
Highly 
appropriate  5  5
300cm; 
118in 5  
Slightly 
appropriate  4 4 4 
Indifferent   3
300cm;
118in 3   3 
300cm;
118in
Slightly 
inappropriate  2  2  2  
Very 
Inappropriate  1  1   1  
  
Given 
Value: 3 3 5 5 3 3
               
  8 5 10 7 8 5
 
Power     
Size of digging apparatus to size of body ratio  
Ratios   Value Known Values 
25-30%  5  
20-24% 30-34%  4 
15-19% 35-39% 3  
10-14% 40-44%  2  
5-9% 45-49% 1 
<5% >50%  0  
  Given Value: 3 0
Applied Force  
Force   Value Known Values 
>85% of its body weight 5  
70-84% of its body weight 4  
55-69% of its body weight 3  
40-54% of its body weight 2  
25-39% of its body weight 1  
<24% of its body weight 0  
  Given Value: 3 0
  109  
Appendix K.  Metric Results for the Townsand Mole 
 
Material Hardness  
Type     Value Known Values 
Rock  5  
Clay   4 
Wood  3  
Compact Soil  2  
Loose Soil  1   
Sand  1  
  Given Value: 1 1
   
  7 1
 
Hard Substances   
Chemical Capabilities  
Capability   Value Known Values 
Yes  5  
Unknown   3 
No  1   
  Given Value: 1 1
Applied Force  
Force   Value Known Values 
>85% of its body weight 5  
70-84% of its body weight 4  
55-69% of its body weight 3  
40-54% of its body weight 2  
25-39% of its body weight 1  
<24% of its body weight 0  
  Given Value: 3 0
Material Hardness  
Type     Value Known Values 
Rock  5  
Clay  4 
Wood   3 
Compact Soil  2  
Loose Soil  1   
Sand  
  Given Value: 1 1
   
5 2
 
  110  
Appendix K.  Metric Results for the Townsand Mole 
 
Mechanical Complexity*   
Motion complexity while digging  
Complexity level  Value Known Values 
Five or more movements 1  
Four movements  2  
Three movements  3  
Two movements  4  
Single movement  5   
List of movements:   
  
Swim motion 
  
  Given Value: 5 5
Material Processing  
Type   Value Known Values 
Compaction  1  
Unknow  3 
Removal  5   
  Given Value: 5 5
Chemical Capabilities  
Capability   Value Known Values 
Yes  1  
Unknown   3 
No  5   
  Given Value: 5 5
   
15 15
  
   
* Values are inverted:  A high mechanical complexity is unfavorable 
 
Geographic Range   
Appropriate Level  Value Known Values 
Highly appropriate  5   
Slightly appropriate  4  
Indifferent  3 
Slightly inappropriate  2  
Very Inappropriate  1  
  Given Value: 5 5
 
5 5
 
  111  
Appendix K.  Metric Results for the Townsand Mole 
 
Size of body   Scenario 1 Scenario 2 Scenario 3 
Appropriate 
Level  Value 
Known 
Values Value 
Known 
Values Value 
Known 
Values 
Highly 
appropriate  5  5  5  
Slightly 
appropriate  4 21cm;8in 4 21cm;8in 4  
Indifferent   3  3  3  
Slightly 
inappropriate  2 2 2 21cm;8in 
Very 
Inappropriate  1 1 1 
  
Given 
Value: 4 4 4 4 2 2
             
  4 4 4 4 2 2
 
Total Value       
 Scenario 1 Scenario 2 Scenario 3 
Speed 4.25 1.25 2.55 0.75 1.7 0.5
Depth  0.4 0.25 1.5 1.05 1.6 1
Power 0.35 0.05 0.7 0.1 2.1 0.3
Hard Substances 0.5 0.2 0.5 0.2 0.5 0.2
Mechanical Complexity 0.75 2.25 0.75 2.25 1.5 1.5
Size of Body  1.2 1.2 1 1 0.3 0.3
Geographic Range 0.5 0.5 0.5 0.5 0.25 0.25
Total Value 7.95 5.7 7.5 5.85 7.95 4.05
       
Raw score with arbitrary values       
    
Raw Score without arbitrary 
values   
    
Weighted = Arbitrary values 
included   
Weighted = Only known values       
    
Information Unknown = 
Arbitrary value of 3       
Information Known     
Data Input