Multi-target Tracking Based on Gaussian Mixture Labeled Multi-Bernoulli Filter with Adaptive Gating

Abstract

In this paper, an adaptive gating technique is applied to labeled multi-Bernoulli filter. In target tracking systems, elliptical gating based on chi-square test is widely used to filter out clutter. However, tracking performance and robustness can be improved by adaptively changing the gate size. To apply adaptive gating to labeled multi-Bernoulli (LMB) filter based on Gaussian mixture, gate size is adjusted by weight of Gaussian mixture. To verify the effect of adaptive gating in LMB filter, simulations for multi-target tracking in cluttered environments are carried out. The simulation results show that the adaptive gating performs similar or better than elliptical gating in terms of OSPA error.