NOVEL GESTURES RECOGNITION AND CLASSIFICATION TECHNIQUE BASED ON EXTRACTION THE INFORMATION FEATURES FROM BISPECTRAL ESTIMATE OF RADIO SIGNAL

Abstract

Novel technique dedicated for human gesture recognition and classification in electromagnetic field of radiofrequency radiation is suggested. Proposed technique is based on extraction of information features by using bispectrum estimation evaluated for high-frequency signal envelope with varied parameters caused by influence of gesture and performed in the interference electromagnetic field in closed space. Robustness performance for information features is experimentally studied by using designed algorithm and software. Obtained results demonstrate that performance of suggested technique depends on signal magnitude and signal delay variations. Technique is sensitive to the signal shape however it largely depends on the signal duration.

Keywords

• радіовипромінювання, обвідна високочастотного радіосигналу, оцінка біспектру, система розпізнавання і класифікації жестів