SIGNAL DETECTORS FOR RANDOM OCEAN MEDIA

This paper reports on acoustic signal detection in a stationary random multipath environment. The multipath transmission channel is modeled considering that both multipath time delay and attenuation coefficients characterizing the emitter/receiver transfer function are random variables with an a priori given distribution. Under the above condition, and assuming the signal-to-noise ratio (SNR) is either low or high, the likelihood ratio (LR) detector structure is developed, analyzed, and interpreted. A Monte Carlo study is also carried out. For low SNR conditions, the statistical behavior of the achieved processor is evaluated and compared to that of two classical detectors: (i) the standard detector derived based on a presumed known multipath channel structure, and (ii) the ad hoc detector developed for inaccurate multipath time delay modeling assumptions.