Adaptive asymptotically robust invariant detection of signals in uncertain noise environments

New adaptive asymptotically robust invariant (AARI) detection algorithm is proposed. The algorithm is used for the detection of signals with some unknown parameters observed in additive noise with unknown distribution. It is based on asymptotically robust invariant (ARI) algorithms previously developed by the authors. The extended model of approximately finite probability distributions (q-point model) is employed for detection algorithm synthesis. Adaptation is achieved by estimating and optimizing the algorithm for the actually unknown parameters of the q-point model. It is demonstrated that, for the case of heavy-tailed noise distributions in DS-CDMA systems, the AARI detector based on the extended q-point model outperforms significantly the correlating detector. The main advantage of the proposed algorithm is its feature of self-adjustment to unknown signal parameters and noise distribution.