AN ADAPTIVE ARRAY DETECTOR WITH MISMATCHED SIGNAL REJECTION

Presented and analyzed here is an adaptive detection algorithm with a sensitivity parameter for rejecting unwanted signals, such as undesired signals entering the sidelobes of a radar or active sonar array. The detection algorithm considered here is a simple modification of the generalized likelihood ratio (GLR) detector (or test) for detecting a signal in zero mean Gaussian noise with unknown correlation matrix. Specifically, the new adaptive detection algorithm is obtained by introducing an arbitrary positive scalar, which we can the sensitivity parameter, into the GLR detector as a multiplier of an already existing quadratic term. The GLR detector then becomes a special case of this detector for unity sensitivity parameter. It is shown that the sensitivity parameter controls the degree to which unwanted signals are rejected. This control of the sensitivity may be important in a surveillance system for example, since as is already known the GLR detector can suffer a relatively large detection loss for signals that are slightly mismatched from the desired signal. From numerical examples, we demonstrate how the sensitivity parameter can be chosen such that unwanted signals, such as sidelobe signals, can be rejected while maintaining acceptable detection loss for slightly mismatched signals, such as mainlobe signals. We also give further insight into previous work on adaptive detection.