Joint statistical signal detection and estimation. Part II: a high-performance closed-loop technique

Abstract

The problem of signal detection and parameter estimation is usually addressed by using separate algorithms; in such a way, the soft information that the detector is possibly capable to provide cannot be exploited by the estimator, and vice versa. This lack of interaction implies a performance degradation, which is particularly appreciable in very low SNR conditions. In this paper we propose a new scheme, which consists of a closed-loop detector and estimator; the key idea is that the results of the estimation algorithm can be used to refine the a priori information at the input of the detection block so as to increase its reliability, and the results of the detection block may be exploited by purposely designed estimation algorithms in order to achieve more and more precise estimates of the parameters (even if this latter feature has not been implemented yet). Such a feedback-aided technique is able to provide a remarkable robustness towards both the noise and the estimation errors, and is even capable to correct decisions that would have been wrong with an open-loop approach. Therefore, it is particularly suitable for those applications which must deal with very low SNR, such as deep-space communications.