Labeled box-particle PHD filter for multi-target tracking
Zhibin Zou, Liping Song, Zhi-long Song · 2017
In multi-target tracking, complex data association problem can be avoided by box-particle filter (box-PF, BPF) based on random finite set (RFS), which reaches similar accuracy results with much considerably less computational costs compared to standard particle filter. However, the BPF based on RFS does not provide identities of individual target state estimates. In this paper, labeled box-particle filter (labeled box-PF filter, LBP) is proposed and its realization for PHD filter is developed in detail, which add a label to each box-particle to record the target identity. It can achieve track management while filtering. The effectiveness and reliability of the proposed algorithm are verified by the simulation results.