Gating technique for the Gaussian mixture multi-Bernoulli filter

Abstract

The multi-Bernoulli (MB) filter is a new attractive approach for multi-target filtering in the presence of clutter and detection uncertainty. However, the computational complexity grows as clutter density increases. The clutter measurements may also degrade the filtering accuracy. To eliminate the clutter measurements and reduce the computational complexity, gating technique for the Gaussian mixture MB (GM-MB) filter is proposed in this paper. Gating technique is designed for each hypothesized track to reduce the number of measurements by discarding most clutter measurements. Simulation results demonstrate the proposed approach can improve the real-time performance without the degradation of filtering performance.