Syntactic classification of radar measurements of commercial aircraft

Abstract

A syntactic pattern recognition system for applications to radar signal identification is designed, and the performance of the resulting classification system is evaluated. Two different techniques for generating symbolic patterns from radar backscatter signals are considered. The classification utility of the symbolic patterns is assessed in terms of the performance of maximum-likelihood classification of the observed symbol strings. A syntax analysis algorithm that makes use of symbolic patterns derived from radar backscatter measurements is developed from the likelihood function classifier. Performance results obtained from simulated classification experiments for both maximum-likelihood and language-theoretic classifiers are presented.<>