Distributed multi-sensor multi-target tracking with feedback

Abstract

An algorithm that incorporates feedback in distributed fusion architectures for maintaining target tracks in cluttered environments is proposed. The decorrelated feedback sequences are constructed by compensating global updated estimates with track information due to global predicted estimates and local track estimates. Due to its orthogonal properties, these feedback sequences are then used in the filtering process to update local estimates before local processors acquire new sets of measurements. The process of constructing these feedback sequences is presented and implemented on a proposed distributed fusion system where each local processor receives measurements from multiple sensors.