Near-field direction finding with a fuzzy neural network

Abstract

The single-source near-field direction finding problem can be solved by a fuzzy nerual network (FNN). The FNN approach can be applied to arrays with arbitrary configurations. It can also be implemented for real-time tracking applications. The approach outperforms the far-field approximation (FFA) approach when the array is uniformly-spaced and Linear, especially when the angle between the array normal direction and the source direction is large and the distance from array center to the source is short.