Detection of GNSS Faults Using Receiver Clock Drift Estimates

Abstract

GNSS based navigation systems are used in almost every sector, from industrial and agricultural environments to everyday applications such as personal navigation or autonomous vehicles. Satellite navigation provides accurate position estimates for these applications, however the performance can be interrupted or degraded by intentional or unintentional interference from inauthentic transmitters. These transmitters can range from GNSS repeaters used for indoor or underground navigation to individuals attempting to fool cellular games into believing they are in another location. This thesis develops a GNSS interference detection method for moving platforms using clock drift estimate monitoring by combining GNSS measurements with inertial navigation systems. During normal operation, GNSS satellites have negligible clock drift, allowing the receiver to estimate the drift of the local oscillator present in the user hardware. When additional transmitters using unsynchronized or low quality oscillators transmit counterfeit GNSS signals, the estimated clock drift will reflect not only the receiver clock drift but also the drift of the transmitter clock. Prior work has demonstrated that estimates of receiver clock drift provide a viable option for integrity monitoring at a static location, however the assumptions shown do not hold for dynamic platforms. By incorporating additional sensors into the algorithm, the clock drift can be estimated for dynamic platforms such as autonomous vehicles or commercial drones. The proposed clock drift monitoring algorithm is analyzed using various quality receiver clocks and inertial sensors, expanding the capabilities beyond that of a standard GNSS receiver. The developed algorithm utilizes time-differenced carrier phase measurements and mechanized INS measurements to provide estimates of the receiver clock drift prior to using the measurements to update the navigation solution, allowing for the detection of measurement faults before the faults influence the state estimate. Two methods of GNSS/INS coupling are compared, which are used to provide position estimates for use in the clock drift estimation algorithm.