A merging approach of the demodulation and equalization for MAP block detection in CPM-SC-FDMA systems

摘要

CPM-SC-FDMA has been recently suggested as an attractive alternative to SC-FDMA multiple access technique which inherits the benefits of the continuous phase modulation (CPM). In this paper, we propose an optimal block detection algorithm for CPM-SC-FDMA systems that merges the frequency domain equalization (FDE), the temporal sample recovery using the inverse DFT (discrete Fourier transform) and the successive CPM decoding operations. The optimal detector derives the MAP (maximum a posteriori) estimate of JT_s-length blocks of symbols by developing a state model description of the questioned problem and then applying a Bayesian detection algorithm referred as to the MAPBD (MAP block detector) algorithm. The MAPBD algorithm is actually a recursive implementation of well-known Bayes filtering equations updating the APP (a posteriori probabilities) vector related to a block of delayed symbols. Simulations show a significant improvement of the BER performance of the MAPBD receiver compared to the conventional CPM-SC-FDMA receiver using an MMSE-FDE in conjunction with a MAP symbol-by-symbol CPM decoder. For complexity requirements, a decision feedback (DF) mechanism is associated to the so-developed MAPBD algorithm as well as a pruning procedure of the APP vector.