An explicit track continuity algorithm based on the GM-PHD filter with adaptive birth intensity

Due to its low computational cost, the Gaussian mixture (GM) implementation of the probability hypothesis density (PHD) filter is a good candidate for multi-target tracking (MTT), especially in real-time processing applications. In this study, we propose an explicit track continuity algorithm based on the GM-PHD filter with unknown target birth intensity for online MTT. In this approach, the identity labels of Gaussian components are divided into three classes, and the labels of confirmed components are unchanged through the overall filtering process, this can provide explicit MTT. Based on the measurements not associated with the predicted components, the adaptive target birth intensity is obtained. The simulation results demonstrate that the proposed approach can achieve explicit MTT under the unknown target birth intensity scenes.