Application of gamma-fluctuating-intensity signal model to active sonar

Sonar performance prediction and detection analysis rely on statistical models for noise, reverberation and signal. Often the central limit theorem is exploited to simplify the modeling. For example, noise and reverberation can be represented by a Gaussian distribution (i.e., Rayleigh-distributed envelope) when they arise from a multitude of sources. Signal models are often limited to the extremes of fluctuating and non-fluctuating. In this paper, a gamma-fluctuating-intensity (GFI) signal model having genesis in the radar research community is applied to sonar to represent variations observed in echoes from an air-filled hose. To facilitate modeling and analysis, exact and approximate techniques to easily evaluate the probability of detection for a GFI signal in Gaussian noise are presented and evaluated. They are then used to examine the impact of the GFI signal on detection threshold and detector non-linearities.