Maximum likelihood estimation for direction of arrival using a nonlinear optimising neural network
T.M. Jelonek, James P. Reilly · 1990
A neural network-based system for a maximum-likelihood estimation of directions of arrival is described. A novel analog neural network implementation of the maximum-likelihood algorithm is presented. Properties of the neural network are discussed with respect to stability and convergence. The performance and behavioral simulations of the network's dynamics are presented. The results show significant improvement over traditional signal-estimation algorithms, such as MUSIC