Consensus-Based Labeled Multi-Bernoulli Filter With Adaptive Event-Triggered Communication

Abstract

The existing event-triggered distributed multi-target filters utilize a constant triggering threshold. Although the communication burden is reduced, it does not adapt well to dynamic environments. In this paper, a novel consensus-based labeled multi-Bernoulli (LMB) filter with event-triggered strategy is introduced with the triggering threshold being decided by a bounded threshold function. The theoretical analysis proves that the information discrepancy of the proposed algorithm is bounded in Kullback-Leibler (KL) divergence sense. The performance of the proposed algorithms is demonstrated in a distributed multi-target tracking scenario via numerical simulations.