A Noise-Assisted Polar Code Attempt Decoding Algorithm

Abstract

Aiming at the characteristics of Polar codes, this paper proposes a noise-assisted Polar code attempt decoding algorithm based on the analysis of frame error rate and bit error rate to improve Polar code decoding performance. The basic principle of the algorithm is that when the encoded data frame is transmitted through the channel, the signal-to-noise ratio is decreased. When the received signal exceeds the decoding capability and cannot be decoded normally, the auxiliary man-made noise is added to make the received signal to have a lower decision probability. The addition of low-power man-made noise will result in further randomization of the region with poor reliability of the original sampled data signal (close to zeros for BPSK modulation), and the bit rollover probability of the higher reliability of the sampled data is smaller. Under this condition, attempt to decode, if check data correct, decode finished, otherwise regenerate man-made noise for the next attempt until the decoding passes or reaches the maximum. This algorithm is a post-compensation processing method for existing algorithms, which play an incremental role.