Estimation of Log-Likelihood Using Constrained Markov-Chain Monte Carlo Simulation

摘要

Log likelihood ratios for data bits transmitted in a multi-dimensional signal are estimated using multiple Markov chain Monte Carlo simulations (MCMC). The MCMC simulations can include constraining symbols based on a most-likely symbol to improve the likelihood of finding distances for non-most-likely symbols. The log likelihood ratios can be calculated based on distances of the most-likely symbol and the non-most-likely symbols.