Robust sequential m-interval approximation detectors with Q-dependent sampling

The theory of a robust sequential detection scheme is extended to include operation with Q-dependent samples of unspecified dependence. The detectors are designed to exhibit a near-optimum performance for nominal noise distributions and maintain high efficiency to changing noise environments by adapting their optimum nonlinearity using an m-interval polynomial approximation (MIPA) of it. The performance of a sequential MIPA detector with Q-dependent sampling is analyzed in detail for two important classes of detection problems: the shift-of-mean alternatives which finds applications in many data communications and image processing problems, and the change-of-scale alternative, which finds applications in processing signal data or spectral estimation data from multisensor arrays