Syntactic pattern recognition of sampled data signals

Statistical pattern recognition techniques perform measurements on sets of data samples to classify pattern types by clustering techniques. Syntactic techniques use a descriptive grammar to see whether the samples fall within an expected shape or envelope. The grammatical representation is described and a comparison provided between the statistical and syntactic techniques. The syntactic techniques is shown to exhibit superior noise tolerance combined with an ability to accommodate various scale factors in signal amplitude and time duration within a single compact grammatical representation.