Analysis of Wideband Beamformers Designed with Artificial Neural Networks

摘要

The ability to determine the direction of an approaching wave is demonstrated byusing an artificial neural network and a beamformer array. To demonstrate the ability of a neural network to learn a satisfactory solution to this problem, simulations were performed to test the system's sensitivity to several variables. These variables include amplitude range, noise level, frequency bandwidth, linear and nonlinear inputs, the number of sensor inputs, and the number of hidden units used in the network. Simulations are provided for both narrow band and wideband signals. An empirical test and a comparison between ANN beamformers and FFT beamformers are also used to demonstrate the strengths and weaknesses of the system. A design example and a description of the simulation program are also included. A brief tutorial on beamformers and neural networks is also provided. (Author)