Distributed joint target detection, tracking and classification via Bernoulli filter

This paper aims to solve the problem of distributed joint detection, tracking and classification (D-JDTC) of a target on a peer-to-peer sensor network. The target can be present or not, can belong to different classes, and depending on its class can behave according to different kinematic modes. Accordingly, it is modelled as a suitably extended Bernoulli random finite set (RFS) uniquely characterized by existence, classification, class-conditioned mode and class & mode-conditioned state probability distributions. Existing algorithms have been devised to perform target JDTC based on a single sensor and can only be easily extended to multiple sensors in a centralized configuration, wherein a fusion centre gathers measurements from all sensors. In this paper, by designing a suitable rule for fusing local posteriors that convey information on target existence, class, mode and state from different sensor nodes, a novel scalable and fault-tolerant D-JDTC Bernoulli filter is proposed, and its performance is evaluated by means of simulation experiments.