T. C. Yang
Naval Res. Lab., Washington, DC 20375
An algorithm is presented for detecting and localizing a passive broadband signal in an underwater acoustic waveguide. The processor estimates the modal spectrum of a broadband signal for source localization; as a consequence, the source spectrum and signature (time domain waveform) are estimated. (Broadband modal decomposition is the equivalent of time domain deconvolution for source signature extraction.) In addition to the usual lofargram, a frequency-depth plot is proposed to display a broadband signal. Source localization is conducted in the frequency domain by summing the (range and depth) ambiguity functions over the signal bandwidth. Sidelobes are suppressed in broadband processing if the signal has sufficient bandwidth. Coherent summation of the ambiguity functions yields the expected time-bandwidth gain if source signature is known (the extended matched filter processing). Numerical simulations are shown for an Arctic waveguide. The processor works for transient as well as continuous broadband signals.