Some new algorithms for multiple maneuvering target tracking

Target tracking is the processing of noise-corrupted measurements obtained from a target in order to maintain an estimate of its current state. In this dissertation we develop a set of new suboptimal filtering and smoothing algorithms for maneuvering target tracking application. The proposed algorithms provide better performance in terms of estimation accuracy over the existing algorithms.; The dissertation is organized in five parts: (1) We propose a novel suboptimal filtering algorithm for the problem of tracking multiple maneuvering targets in the presence of clutter. In the proposed algorithm, we exploit the multiscan joint probabilistic data association (Mscan-JPDA) approach to address the measurement-to-track association problem. Earlier work on the multiscan JPDA is restricted to non-maneuvering targets, in which single kinematic model is used to model the motion behavior of the target. We use switching multiple model approach in which distinct kinematic models describe the motion of the target and switching among these models is governed by a Markov chain. We then exploit the basic interacting multiple model (IMM) approach for multiple model estimation. (2) We extend the proposed IMM/Mscan-JPDA filtering algorithm to the fixed-lag smoothing case in which the objective is to obtain the state estimates at time k when given the measurements up to time k + d ( d > 0 and is fixed). The smoothing algorithm achieves significant improvement in the estimation accuracy over the filtering algorithm, by introducing a small time lag between the instants of estimation and latest measurements. (3) We extend the proposed IMM/Mscan-JPDA filtering algorithm to a multisensor tracking scenario. We carry out sequential updating of the state by obtaining measurements from the multiple sensors. (4) We consider the problem of scheduling multiple sensors for a maneuvering target tracking application. Tracking of highly maneuvering target is achieved by an effective suboptimal filtering algorithm based on the IMM filtering approach combined with the probabilistic data association (PDA) technique and the proposed sensor scheduling algorithm. (5) We propose a new adaptive sampling policy for a maneuvering target tracking application. Multisensor tracking is achieved by a suboptimal filtering algorithm developed by the IMM/PDA filter combined with the proposed adaptive sampling scheme. (Abstract shortened by UMI.)...