Journals Proceedings

Abstract

In this paper, we developed a new State-Drop Fast Sequence Estimation (SDFSE) strategy for Trellis Coded Modulation (TCM) schemes for transmission in the Intersymbol Interference (ISI) environment. For the decoding of TCM signals in the presence of Additive White Gaussian Noise (AWGN), Maximum Likelihood Sequence Estimation (MLSE) has been considered as the optimum solution [13,16,17,19]. However, for band-limited ISI channels in the presence of AWGN, the complexity of optimum MLSE increases as a function of the ISI channel memory length. This prohibits practical implementation of MLSE. Over the past decades, a spurious research took place in the development of reduced complexity suboptimum decoding strategies for TCM schemes. Reduced State Sequence Estimation (RSSE) is one such implementation which emphasizes on reduced computational complexity Likelihood sequence estimation by minimizing the ISI-code trellis states [23]. We provide a new suboptimum decoding strategy, a reduced computational complexity SDFSE which takes the path metrics of soft output Viterbi algorithm as a measure to decide and compute the state transitions of only-best-survivor in the succeeding intervals. The SDFSE results in a reduced number of nodes expansions during Likelihood sequence estimation as compared to RSSE. A decision parameter in comparison with the accumulated path metric of the best survivor is tuned to provide variable complexity for the algorithm. We evaluated the error performance of SDFSE through computer simulation for 4-state 16-QAM TCM scheme in the ISI environment. The results are compared with the error performance of RSSE, which we consider as conventional-RSSE (c-RSSE). It is found that computational complexity of SDFSE is less, and hence is faster than c-RSSE. The SDFSE provides an error performance close to c-RSSE, and is a function of the decision parameter of SDFSE strategy.