Fiducial Marker-Based Motion Capture System for Biomechanics and Lower Limb Prosthesis Alignment

Abstract

Lower limb loss is a disability that is currently estimated to affect more than one million Americans. Development of lower limb prostheses has allowed the amputee to regain their mobility to some extent and reengage in society, work, sports, military service, and other activities. However, correct alignment of the prosthetic limb is even more critical as it may lead to discomfort, pain and serious diseases such as osteoarthritis, osteopenia, and osteoporosis. Currently, prosthetic alignment is carried out mostly by eye-balling or with one of the static alignment devices available in the market. Additionally, lower limb prosthetic alignment is an iterative and time-consuming procedure that can be emotionally difficult and frustrating for the patient as it requires multiple appointments, fitting trials, and alignment events. Thus, the goal of this thesis is to make developments towards a portable system that can deliver real-time alignment data to the prosthetist resulting in a more expedient prosthetic alignment process. This will allow the prosthetist to focus more on patient care and less on constant adjustment.

In this study, a cost effective, self-contained, and portable system for prosthetic alignment was developed utilizing a low-cost fiducial marker tracking system, which also included determination of the landmarks to be tracked. The identified landmarks were then used to develop algorithms to calculate the joint angle kinematics for each lower limb prosthetic joint. The system also includes a webpage application with the ability to digitally provide biomechanical (static and dynamic) alignment data to the prosthetists. A pilot validation study utilizing conventional motion capture system as gold standard was conducted in order to validate the fiducial marker system for use in human motion capture and prostheses alignment. The study yielded a RMSE of 3.3±0.91˚for flexion/extension angle of the prosthetic limb with an average difference in maximum flexion and extension angles between the systems of 0.035±0.9˚ and 1.96±0.04˚, respectively.

The second goal of the thesis study was to develop a component-based coordinate system for lower limb prosthesis. The technique presented in this study provides a method of creating a component-specific coordinate system for a standard transtibial prosthesis. The developed coordinate system was used for the purpose of calculating objective measures of alignment while the prosthesis was being worn during both static and dynamic activities. The development method is based on the selection of anatomically relevant points tracked on the residual limb and the prosthesis along with Euler angle decomposition techniques. No such method has been developed previously for a transtibial prosthesis.

The third goal of the study included validation of the augmented reality marker-based tracking system for utilization in general motion capture, specifically for prosthetic alignment. The proposed study was validated on five physically fit participants of varying age using a traditional retroreflective marker-based motion capture system. This study design tested the effect on position and orientation of fiducial markers of 3 different marker sizes placed on the thigh of participants walking at 3 different speeds walking from 0 to 3.5 m away from the action camera. The results indicated an increase in RMSE for the position and orientation of the fiducial marker with an increase in speed and decrease in fiducial marker size. The results were comparable to the previous studies performed at static condition of fiducial markers for utilization in augmented reality. Thus, the results obtained in this study suggest that the fiducial marker system can be used for general motion capture and lower limb prosthesis alignment.