Online Spatial–Temporal Alignment Method in Bearing-Only Sensors for Maneuvering Target Tracking

Abstract

Spatial and temporal alignments are crucial preprocessing procedures in multisensor tracking systems. Earlier studies rarely considered the alignment aspect in target tracking. This article proposes an online spatial-temporal alignment (OSTA) method in the framework of the interacting multiple model (IMM) estimator and square-root cubature Kalman filter (SRCKF) for maneuvering target tracking. Based on motion models and density clustering, an online temporal alignment (OTA) method is proposed to smoothly process the measurements of each sensor received in a fusion period. Combined with the OTA method, online spatial alignment is considered as the augmented state (AS) of the target to be jointly estimated by the AS Kalman filter (ASKF). To handle filter initialization in bearing-only sensors, the initial AS and its covariance are derived using the one-point initialization method. The IMM estimator is incorporated with the SRCKF to achieve a joint estimation of spatial alignment and maneuvering target tracking. The posterior Cramer-Rao lower bound (PCRLB) is used to evaluate the estimated performance. Numerical simulations are performed to demonstrate the effectiveness and superiority of the proposed IMM-SRCKF–OSTA method.