An evolutionary computing approach for the target motion analysis (TMA) problem for underwater tracks

This study is concerned with the problem of determination of kinematic parameters of a moving sound source by using only bearings known as target motion analysis (TMA). A new matched field signal processing approach which uses genetic algorithm (GA) and Monte Carlo simulation is proposed to establish tracks from bearing-only contacts. The basic idea is to take a number of measurements and run a simulation of underwater tactical situation, then to let the simulation change its parameters until the output matches with the measurement data. When the simulation output (i.e. the replica data) matches the real measurement within a predefined degree, it is expected that the simulation resembles the real situation. In this sense, the TMA problem is considered as an optimization problem on a large parameter space and genetic algorithm was used to solve it. We developed an application called target motion analysis with genetic algorithm (TMAGA) in order to demonstrate the correctness of the algorithm. Monte Carlo simulations demonstrate the results of the experiments conducted with TMAGA.