Range and bearing estimation under occlusion using variants of state-space recursive least squares

In practical trajectory estimation problem, occlusion (missing observation) degrades estimator precision and more seriously, may result in estimator divergence. In this paper, we present a modification to state-space recursive least squares (SSRLS) for trajectory estimation in the presence of observation noise and occlusion. By incorporating the inherent extrapolation feature of SSRLS, a dual-mode form of SSRLS is proposed for occluded trajectory estimation. This paper provides details of proposed algorithm followed by straight line and maneuvering trajectory estimation examples. Dual-mode SSRLS and its variants, when compared with Recursive Least Squares (RLS), Kalman Filter (KF) and Extended Kalman filter (EKF), exhibit remarkable tracking performance under occlusion. In case of unknown trajectory dynamics, superior tracking abilities of SSRLS are highlighted with the results supporting the validity of our claim.