A low-complexity rate compatible modulation via variable weight sets

Rate Compatible Modulation (RCM) had been shown to be capable of achieving smooth rate adjustment in highly dynamic channel scenarios. However, there still exist two main problems in RCM: (1) The theoretical bound of the achievable rate, which is the benchmark for practical RCM design, has not been obtained. (2) The iterative decoding algorithm for RCM have a high computational complexity due to the long weight set with large value and the much more symbols required by receiver for successful decoding under poor channel conditions. To tackle these problems, Firstly, we get the theoretical achievable rate upper bound of RCM via analysis of the mutual information between transmitted and received symbols, and obtain an important result, i.e., in low SNR, the achievable rates of RCM are completely the same for the cases of large or small weight sets. Based on this result, a low-complexity RCM with variable weight sets (RCM-VWS) is proposed. In addition, we investigate the messages relationship between symbol nodes and information nodes in decoding algorithm based on Bayesian interference and analyze the computational complexity of the RCM-VWS we proposed. Theoretical analysis and simulations show that, in low SNR (<;8dB), the computational complexity of the proposed scheme can be reduced by 95% compared with that of the conventional RCM, while preserving the same BER performance. So the average decoding efficiency is improved significantly.