Robust GPS Geolocation by Method of Particle Filtering

Howard L. Dyckman, Shön Sloat and Bill Pettus

Abstract:	The potential vulnerability of GPS to collateral interference and intentional jamming presents a major concern to both military and civil users. In response, a number of electronic protection (EP) strategies are now successfully deployed to mitigate the impact of GPS interference sources. However, EP strategies such as nulling antennas may also reduce satellite visibility along with interference sources. Also, GPS ground receivers deployed within dense forest or urban areas may encounter physical obstructions that block available GPS satellites, precluding the receiver from obtaining optimum Geometric Dilution of Precision (GDOP) and perhaps even from tracking a minimum number of satellites. Kalman Filtering (KF) techniques and their applications to GPS, INS, and coupled systems are well understood. However, Kalman and Extended Kalman Filters are best suited when the application exhibits linear tendencies and quickly degrade as the application becomes nonlinear. Particle Filtering [1-9] presents a robust alternative to Kalman Filtering that is able to accommodate systems exhibiting arbitrary nonlinearity. Relevant nonlinearities include inertial navigation state equations and the underlying geolocation calculation of a GPS receiver. In contrast to the single solution tracking inherent in Kalman Filter implementation, the Particle Filter tracks multiple candidate solutions ("particles") at various points within the nonlinear solution space. Particle Filtering can compensate for the lack of sufficient instantaneous GPS satellite observability by "statistically stitching" together observations made at different times. The various observations may be of different satellites, or of the same satellite observed at different times (after motion). This paper provides simulation results for cases where too few GPS satellites are simultaneously visible to provide a full GPS solution. Specifically, Particle Filtering is demonstrated to yield a superior solution as compared to the Extended Kalman Filter under circumstances of diminished satellite observability.
Published in:	Proceedings of the 2005 National Technical Meeting of The Institute of Navigation January 24 - 26, 2005 The Catamaran Resort Hotel San Diego, CA
Pages:	197 - 204
Cite this article:	Dyckman, Howard L., Sloat, Shön, Pettus, Bill, "Robust GPS Geolocation by Method of Particle Filtering," Proceedings of the 2005 National Technical Meeting of The Institute of Navigation, San Diego, CA, January 2005, pp. 197-204.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In