Tracking with Sparse and Correlated Measurements via a Shrinkage-based Particle Filter

摘要

This paper presents a shrinkage-based particle filterudmethod for tracking a mobile user in wireless networks. Theudproposed method estimates the shadowing noise covarianceudmatrix using the shrinkage technique. The particle filter isuddesigned with the estimated covariance matrix to improve theudtracking performance. The shrinkage-based particle filter canudbe applied in a number of applications for navigation, trackingudand localization when the available sensor measurements areudcorrelated and sparse. The performance of the shrinkage-basedudparticle filter is compared with the posterior Cramer-Rao lowerudbound, which is also derived in the paper. The advantagesudof the proposed shrinkage-based particle filter approach areuddemonstrated via simulation and experimental results.