MODULAR LEARNING-STRATEGY FOR SIGNAL-DETECTION IN A NONSTATIONARY ENVIRONMENT

In this paper, we describe a novel modular learning strategy for the d etection of a target signal of interest in a nonstationary environment , which is motivated by the information preservation rule, The strateg y makes no assumptions on the environment, It incorporates three funct ional blocks: 1) time-frequency analysis, 2) feature extraction, 3) pa ttern classification the delineations of which are guided by the infor mation preservation rule, The time-frequency analysis, which is implem ented using the Wigner-Ville distribution (WVD), transforms the incomi ng received signal into a time-frequency image that accounts for the t ime-varying nature of the received signal's spectral content, This ima ge provides a common input to a pair of channels, one of which is adap tively matched to the interference acting alone, and the other is adap tively matched to the target signal plus interference, Each channel of the receiver consists of a principal components analyzer (for feature extraction) followed by a multilayer perceptron (for feature classifi cation), which are implemented using self-organized and supervised for ms of learning in feedforward neural networks, respectively, Experimen tal results based on real-life radar data are presented to demonstrate the superior performance of the new detection strategy over a convent ional detector using constant false-alarm rate (CFAR) processing, The data used in the experiment pertain to an ocean environment, represent ing radar returns from small ice targets buried in sea clutter; they w ere collected with an instrument-quality coherent radar and properly g round truthed.