Gaussian particle filtering for tracking maneuvering targets

Abstract

Tracking for maneuvering targets in the presence of clutter is a challenging problem. In this paper, we present an algorithm for reliable tracking of maneuvering targets based on Gaussian particle filtering (GPF) techniques. It has been shown that sequential Monte Carlo (SMC) methods outperform the various Kalman filter based algorithms for nonlinear tracking models. The SMC, also known as particle filtering, methods approximate the posterior probability distribution of the parameter of interest using discrete random measures. GPF is another variant of the SMC methods which approximates the posterior distribution using a single Gaussian filter. Unlike the standard SMC methods GPF does not require particle resampling. This distinct advantage makes GPF to be easily amenable to parallel implementation using VLSI. The proposed tracker is tested in a fairly complex target trajectory. The target maneuvering is simulated using Markov jump process of three kinematics models having different accelerations. Computer simulations show the proposed algorithm exhibits excellent tracking capability in a fairly complex target maneuvering.