Evaluation of the Block Error Probability of Block Modulation Codes by the Maximum-Likelihood Decoding for an Awgn Channel

This paper is concerned with the evaluation of the block error probability P/sub e/ of a block modulation code by optimum or suboptimum soft-decision decoding for an AWGN channel. In the range D/sub L/ of low signal-to-noise ratios, it is feasible to evaluate P/sub e/ by simulation. In the range D/sub H/ of very high signal-to-noise ratios, tight upper bounds on P/sub e/ are available. However, in most cases, there remains a gap between D/sub L/ and D/sub H/, and the value P/sub e/ for signal-to-noise ratios in this gap may be of practical interest. In general, very small values of the block error probability are required. It is infeasible to evaluate a very small P/sub e/ simply by simulation. In this paper, the maximum-likelihood soft-decision decoding of a block modulation code for an AWGN channel is considered, and a new method of evaluating the block error probability P/sub e/ for a wide range of signal-to-noise ratios is presented. The evaluation of P/sub e/ is reduced to numerical computations and simulations on statistics with not very small mean value. Computation results for the block error probabilities of some block modulation codes are given.