Ambiguous measurement update in terrain-referenced navigation with particle filtering

Abstract

Ambiguous measurement is often found near the inclination points of the measurement model where the same measurement pertains to two or more different states. This dissertation addresses the observation that the ambiguous measurement update, which causes increase in estimated covariance, is contributed by not only the shape of the measurement model but also the prior distribution of the filtered estimate. We give a novel suggestion that a way to cope with the local measurement ambiguity is to reshape the prior density, and propose a particle filtering algorithm which adopts an established solution to the out-of-sequence measurement (OOSM) problem on the framework of the particle filter with sequential importance resampling (SIR). This strategy provides a remedy to the ambiguity problem to obtain accurate current estimate with lower covariance. Numerical simulation for terrain-referenced navigation is presented to demonstrate effectiveness and performance of the proposed method. Compared to other methods such as the standard particle filter, the auxiliary particle filter, the mixture particle filter, and the receding-horizon Kalman filter, the proposed method shows better performance in terms of root-mean-square error and estimated covariance.