Adaptive Gaussian Mixture-Probability Hypothesis Density Based Multi Sensor Multi-Target Tracking

This paper addresses a novel multiple target tracking (MTT) problem in a decentralized sensors network (DSN) framework. The algorithm jointly estimates the number of targets and the states of the targets from a noisy measurement in the presence of data association uncertainty and missed detection. The standard GM-PHD filters estimate the multi-targets in a cluttered environment with an assumption that the target birth intensity is known or homogeneous. It results in inefficient tracking for new, occluded or missed targets. The issue is addressed by the proposed adaptive Gaussian birth components based estimation. A method based on covariance intersection fusion is proposed to address inter-sensor target data association.