Estimation and detection in matched field processing

Matched field processing (MFP) is a technique used for processing ocean acoustic information that uses a propagation model in conjunction with measured data. It has historically been treated as an estimation problem, but in the weak signal case the detection problem must be considered. In this work we begin by reviewing the formal estimation structures as evolving from the Bayesian approach which, in turn, lead to the classical or likelihood methods. Next, we show how the techniques that are presently being used in MFP fit into this formal structure. Finally, the joint detection/estimation problem, which is central to MFP, is discussed. In particular, we consider the case where p(H₁ ), the probability of the source being “on,” is known to be less than unity