A generalization of multiple-symbol differential detection and its application to fast time-varying fading channels

This paper discusses a generalized concept of multiple-symbol differential detection (MDD) and analytically derives parameters for MDD on fast time-varying channels. At first, this paper shows that adaptive maximum-likelihood sequence estimation (MLSE) employing the per-survivor processing (PSP-MLSE) with a single channel tap similar to the concept of MDD. Next, parameters for MDD are derived according to the channel estimation of PSP-MLSE based on a high order channel prediction. Finally, computer simulation confirms that MDD with the analytically derived parameters has excellent tracking capability on fast time-varying fading channels.