Blind focusing algorithm applied to the acoustic signal of a maneuvering rotorcraft

An algorithm was developed and tested to blindly focus the acoustic spectra of a rotorcraft that was blurred by time-varying Doppler shifts and other effects such atmospheric distortion. First, the fundamental frequency generated by the main rotor blades of a rotorcraft was tracked using a fixed-lag smoother. Then, the frequency estimates were used to resample the data using interpolation. Next, the motion-compensated data were further focused using a technique based upon the phase gradient autofocus (PGA) algorithm. The performance of the algorithm was evaluated by analyzing the increase in the amplitude of the harmonics due to focusing the data. For most of the data, the algorithm focused the harmonics between approximately 10-90 Hz to within 1-2 dB of an estimated upper bound (UB) obtained from conservation of energy and estimates of the Doppler shift. In addition, the algorithm was able to separate two closely spaced frequencies in the spectra of the rotorcraft. The algorithm developed can be used to preprocess data for classification, nulling, and tracking algorithms.