A Fast Multi?Target Tracking Algorithm Based on Maximum Entropy Fuzzy Clustering
Xiao Chen,Yaan Li,Jing Yu,Yuxing Li
DOI: https://doi.org/10.3969/j.issn.1000-2758.2017.04.011
2017-01-01
Abstract:In order to improve the real?time and the accuracy of tracking multi?target in dense clutter environment,a new data association algorithm based on maximum entropy fuzzy clustering data association was introduced in this paper, which uses fuzzy membership matrix to express the association probability between target and measurement. At the same time, the effect of the public measurement to the target was analyzed, which uses impact factor to re?construct association probability matrix, combining with the probabilistic data association algorithm to estimate the state of target. This algorithm avoids the splitting the confirmation matrix, so it can solve the problem of the high computational load of the joint probabilistic data association algorithm with the increase of the clutter and target. In addition, simulation and analysis are carried out for tracking parallel targets and cross targets in different clutter density environment. Simulation shows that this algorithm is an efficient and fast data association algorithm, and the tracking performance is superior to the joint probabilistic data association algorithm and the cheap joint probabilistic data association algorithm in dense clutter environment, and also can avoiding the tracking coalescence to a certain extent.