부산대학교 도서관

The successive cancellation list (SCL) decoding of polar codes can achieve a performance close to the maximum-likelihood (ML) decoding while large list size results in high computational complexity. To this end, successive cancellation priority (SCP) decoding is proposed to reduce the complexity of the SCL decoding with acceptable space complexity. In this paper, an adaptive SCP (ADSCP) decoder with iteratively increased searching width is proposed. Two priority queues are applied to store all the priority nodes during the decoding. When the correct decoding path cannot be obtained with the current searching width, the decoding continues with a larger searching width rather than be restarted. Furthermore, an early termination criterion is applied to quit the decoding when too many inefficient path extensions are performed based on a preset threshold. The early termination criterion can reduce the worst-case complexity of the decoder. Simulation results show that given the same searching width (list size), the proposed ADSCP decoder can achieve the same decoding performance as the conventional adaptive SCL (ADSCL) decoder while significantly reducing the complexity.