Bio-inspired optimisation approach for data association in target tracking

Data association is an essential part of track maintenance in multiple target tracking, which can be solved by multidimensional assignment methods. When there is a need to solve the multidimensional assignment problem, the ant colony optimisation (ACO) algorithm stands out as it can solve combinatorial optimisation problem with excellent performance in acceptable CPU time. Here, each measurement is modelled as an ant, each track is modelled as a city, and the problem of data association is modelled as the food locating by ants. Thus, a novel data association based on an improved ant colony optimisation algorithm (ACODA) is proposed in this paper. The detailed corresponding relationship and theoretical analysis between basic ACO algorithm and the ACODA algorithm are given. Simulation results show that as the number of targets increases, the ACODA algorithm performs better than JPDA and NN, with superior performance both in computational time and accuracy.