Neural networks for direction finding via a stochastic model

A neural network approach is presented for finding the approximate maximum likelihood estimates of the direction-of-arrival (DOA) of plane waves. Based on the model of the maximum likelihood (ML) estimator, the direction finding problem is mapped onto the Lyapunov energy function of the Hopfield model neural network. To make the network converge to a valid solution at a low SNR value, the information of source numbers is also constrained in the network's energy function. Simulation results are presented to illustrate the improved performance achieved by this new approach.<
>