Learning neural networks for detection and classification of synchronous recurrent transient signals

This paper proposes a neural network solution to the classical signal processing problem of detection of a synchronous recurrent transient signal in noise. If a signal exists, it is assumed to be one of M known signals which may sometimes occur (probabilistically) in successive intervals. Several neural network configurations are applied to this problem and compared with each other and with the optimum adaptive sequential detector. A novel efficient neural network detector is proposed using an XOR-Tree configuration with learning. Tests with synthetic and real noise, show the excellent performance of this approach as compared to the optimum adaptive detector and to other neural network techniques. With real (non-white) noise obtained from sonar data, the XOR-Tree network widely outperforms the likelihood ratio detector. We also discuss the learning time complexity of the XOR-Tree network and compare it to that of standard three layer network architectures. © 1998 Elsevier Science B.V. All rights reserved.

Duke Scholars

Author Jeffrey L. Krolik Electrical and Computer Engineering