Maneuvering Multitarget Tracking in Clutter with the VSMM Estimator and Neural Network

摘要

In this paper, the problem of tracking maneuvering multitarget in clutter is studied. In order to alleviate the complexity and computation burden, a new approach using the neural network and Variable Structure Multiple Model (VSMM) estimator to solve the data association and maneuvering target tracking problems is presented. The neural network used here offers a significant degree of parallelism and thus can compute the probabilities of the various generated hypotheses more rapidly. Meanwhile, instead of solving a large problem, the entire set of targets and measurements is divided into several clusters that a number of smaller problems can be solved independently. Computer simulations have shown that the presented method has high convergence performance, good accuracy and robustness to uncertainty of target and moderately dense clutter.