Evolutionary TDOA-based direction finding methods with 3-D acoustic array

This paper focuses on direction finding of a signal source using time-difference-of-arrival (TDOA) measurements at a 3-D acoustic sensor array. Two evolutionary computation methods are proposed to solve the direction finding problem, which are the genetic algorithm and the particle swarm optimization algorithm. Sound speed is used in the development of the algorithms, which is estimated based on observed weather parameters and initial direction estimation results from the least square (LS) estimator which is presently the key method in TDOA-based direction finding. All reference-free TDOA measurements are adopted in defining cost function to improve performance. To guarantee fast convergence, an LS estimator is also utilized to provide initial direction estimates for the two swarm intelligent algorithms. Simulation results demonstrate that the proposed methods with a full TDOA set are superior to the Cramer-Rao lower bound with a limited set of reference-based TDOA measurement, significantly outperforming the LS estimator. Extensive field experiments were conducted and there is good agreement between the experimental results and simulation results.