A multi-swarm particle swarm optimization algorithm for tracking multiple targets

Abstract

A nove multi-swarm particle swarm optimization algorithm to tracking multiple targets is proposed. The global swarm is divided into many sub-swarms according to the number of targets, and each sub-swarm searches for its target and maintains track of it. The information interaction between different sub-swarm is described as a differentiable energy function, which include an observation model and a mutual exclusion model. Experimental results demonstrate the efficiency of our method.