MINIMAX DETECTION FOR SIGNALS WITH UNKNOWN-PARAMETERS

Minimax detectors are often used for the detection of signals with unk nown parameters in additive white Gaussian noise. Several minimax stra tegies are discussed for both real and complex signals, including one based on the least favorable distribution of the unknown parameters an d a second, distribution-free method based on the worst case parameter values, Also discussed are the relationships between these minimax de tectors and equal correlation filters for real signals and equal defle ction quadratic detectors for complex signals. Conditions are given un der which the two different minimax strategies produce identical detec tors.