Statistics of biological noise and performance of generalized energy detectors for passive detection

Underwater noise due to snapping shrimp is highly impulsive, and often domi nates the ambient noise environment of warm, shallow waters at frequencies above 1 kHz. We report here on the statistics of bandpass snapping shrimp n oise data, and on the modeling of the joint distribution of the in-phase an d quadrature components using bivariate versions of the generalized Gaussia n (GG), generalized Cauchy, and Gaussian-Gaussian mixture models. We evalua te the performance of several generalized energy detectors for passive band pass detection, by inserting stochastic signals into the noise data. Detect ion thresholds were measured for an integration time of 0.5 s and false ala rm probabilities down to 1%. The locally optimum detector based on the mixt ure model gave the best weak signal detection performance, with an 8 dB red uction in detection threshold over conventional energy detection. A signifi cance test detector based on the GG model performed 1-2 dB worse, but exhib ited better strong signal performance.