Guided wave signal recognition by matching pursuit based on evolutionary programming algorithm

Evolutionary programming using mutations based on the t probability distribution (tEP) is introduced and tested by maximizing the test function. The evolutionary programming algorithm is applied to the matching pursuit method. A steel pipe with a notch is tested by guided wave testing system and the measured signal is decomposed by the matching pursuit method. The defective signal is easy to recognize from the processed signal. The defect location analyzed theoretically according to the matched result agrees well with the experiment setup. The processed signal is compared with the processed result obtained by matching pursuit based on DE, and the features are nearly the same, except that there are some difference between the wave amplitude and wave width. Therefore, matching pursuit based on tEP is a useful method to process guided wave signals and to recognize defective signals.