Neural network based optimum radar target detection in non-Gaussian noise

The application of neural networks to radar target detection in non-Gaussian noise environments is investigated. Two new probabilistic neural networks, the Gram-Charlier neural network and the Gram-Charlier probabilistic neural network, were applied to the radar detection. The performance of these detectors was evaluated and compared with backpropagation and Bayesian classifiers by simulation for Gaussian, Weibull, and lognormal noise environments.<>