ADAPTIVE WAVELET SIGNAL PROCESSING

MODERN SENSORS PRODUCE VAST AMOUNTS OF SIGNAL INFORMATION WHICH MUST BE PROCESSED QUICKLY AND ACCURATELY TO PERFORM SURVEILLANCE AND TARGET TRACKING FUNCTIONS. PREPROCESSING OF THE SENSOR DATA IS PERFORMED FIRST TO REMOVE BACKGROUND SIGNALS, DETECT SPIKES AND PEAKS, FILTER THE SIGNAL, ENHANCE EDGES AND IMPROVE THE SIGNAL-TO-NOISE RATIO. FURTHER PROCESSING IS THEN PERFORMED TO IDENTIFY OBJECT DIRECTION, ORIENTATION AND TO DISCRIMINATE THE OBJECT USING RANGE AND VELOCITY ESTIMATIONS, KALMAN FILTERS AND BY OTHER METHODS. IN PHASE I FASTMAN PROPOSES TO DEVELOP A NEW SIGNAL PREPROCESOR BASED ON THE ADAPTIVE WAVELET TRANSFORM. THE ADAPTIVE WAVELET TRANSFORM DECOMPOSES A COMPLEX SIGNAL INTO ITS INTEGRAL COMPONENTS ALLOWING KEY SIGNAL FEATURES TO BE READILY EXTRACTED. WE BELIEVE THAT THE ADAPTIVE WAVELET TRANSFORM IS CAPABLE OF ENABLING ORDER-OF-MAGNITUDE IMPROVEMENTS IN SIGNAL PROCESSING PERFORMANCE. IN THIS PROJECT WE WILL: 1) ESTABLISH A MATHEMATICAL BASIS FOR USING THE ADAPTIVE WAVELET TRANSFORM TO REPRESENT COMPLEX SIGNALS, AND 2) DEVELOP STRATEGIES TO EXTRACT TARGET SIGNATURES AND KEY SIGNAL FEATURES FROM THE SIGNAL REPRESENTATION OBTAINED THROUGH THE ADAPTIVE WAVELET TRANSFORM.