CRC-Aided Adaptive Successive Cancellation List Decoder for 5G NR Polar Codes Using RBIR

Cyclic redundancy check (CRC)-aided successive cancellation list (SCL) decoder, also termed as CA-SCL decoder, is used for decoding the polar codes in practical 5G new radio (NR) systems. The list size used in CA-SCL decoder is directly proportional to the power consumption and the complexity of the decoding process. In practice, a fixed list size (L) is used for CA-SCL decoding. However, we have observed that the required list size varies with signal to noise ratio (SNR) for achieving a target block error rate (BLER) performance. In this paper, we propose an adaptive CA-SCL (CA-ASCL) decoding algorithm, that reduces the average list size and hence the complexity of the CA-SCL decoder by identifying the right list size to be used for each decoding process, using a link abstraction metric called received bit information rate (RBIR), that maps to received SNR. We show that for a target BLER of 10%, the proposed CA-ASCL decoder that chooses the list size adaptively among the set {2, 4, 8, 16, 32} has upto 61.51% reduction in list size compared to that of CA-SCL decoder with L = 32, without any loss in BLER performance, and upto 72.23% reduction with performance loss limited to 0.05 dB in SNR, respectively, for 3GPP defined 5G NR TDL-A channel model, with delay spread of 30 ns and maximum Doppler shift of 100 Hz.