Improvement of AR Code Decoding Algorithm Based on the LLR Information Value

Rateless code is applied to the wireless communication system, to solve the rate mismatch problem of fixed rate code due to the channel change. There is a good compromise of Accumulate Rateless Code (AR code) in rateless family on the decoding performance and complexity. But Log Likelihood Ratio-Belief Propagation (LLR-BP) decoding algorithm is based on the large consumption of time and resources. In this paper, to improve the decoding efficiency of AR code, we proposed an improved decoding algorithm based on LLR information value. By means of tracking the LLR information value of AR code in the process of decoding, and researching on the probability distribution of right and wrong decode symbol information value, we gave the best selection strategy of value_LLR threshold. The simulation analysis show that the proposed algorithm compared with traditional reset and not reset algorithms in decoding performance keeps almost the same, but it reduces 5.9% and 14.0% of the time consumption respectively.