Ant estimator with application to target tracking

Standard ant colony optimization (ACO) algorithm is usually utilized to solve various hard combinatorial optimization problems. In this paper, however, the idea of the generic ACO is extended to the scope of recursive parameter estimation, i.e., ant estimator is first proposed and investigated to track target of interest. In the proposed ant estimator framework, some basic properties of particle filter (PF) and ant colony optimization (ACO) are inherited, and the “fittest variables” are determined by ants with probability decisions. In addition, the pheromone update strategy is also well defined in order to guide more ants to better solutions. Finally, two improved versions of the original ant estimator are investigated and applied to some benchmark target tracking problems. Numerical experiments demonstrate that these proposed ant estimators perform well, and moreover, each could deal with maneuvering target tracking problem without any auxiliary technique.