Research On Concatenated Polar Codes For 5G Communication System

Polar codes are the first class of error correction codes which have been proved to achieve the symmetric capacity of binary-input discrete memoryless channels(B-DMCs).The error correction performance of polar codes is significantly improved by the advanced decoding algorithms and concatenated polar codes including cyclic redundancy check polar(CRCPolar)codes,parity check polar(PC-Polar)codes,and Hash-polar codes,which made it comparable to low density parity check(LDPC)codes,turbo codes and other state-of-theart channel codes.Considering the evolution of agreements on concatenated polar codes in control channel of 5G enhanced mobile broadband(e MBB)scenario and the new requirements of ultra reliable low latency communications(u RLLC)scenario,we study the concatenated polar codes in 5G communication system.The main work can be summarized as follows.1.The fundamentals of polar codes and several concatenated polar codes are introduced.The performance of different construction methods,decoding algorithms,rate matching solutions and concatenated polar codes are also given.2.The software simulation platform of control channel in 5G e MBB scenario at physical level is built and the performance of uplink and downlink control channels are provided.A group parallel interleaver scheme is proposed to improve parallelism degree of the existing triangle interleaver scheme.We show that the error correction performance of the group parallel interleaver scheme is not degraded.A segmented decoding strategy is proposed to improve the performance of long concatenated polar codes at the cost of a small amount of extra computation under the agreed rate matching scheme.3.Due to the strict requirement of false alarm ratio(FAR)and the demand of early termination(ET)for decoder in e MBB scenario,a kind of distributed CRC-polar(DCRC-Polar)codes based on the restrict structure is proposed which can satisfy the FAR requirements.We extend the restrict structure to analyse the influence factors of FAR and ET gains of DCRCpolar codes and give a method about how to design FAR-satisfied DCRC polar codes with high ET gains.This method is also suitable for segmented Hash-polar codes.The simulation results show the segmented Hash-polar codes achieve higher ET gains than the agreed DCRC-polar codes.4.Since low decoding latency is required for 5G u RLLC scenario,we study the impact of construction on the performance of belief propagation list(BPL)decoding algorithm of polar codes.The BPL decoding performance under different code constructions is given.And we apply it to the decoding of LDPC-Polar codes,which improves the BLER performance further.A hybird decoding strategy of LDPC-Polar codes is also proposed.According to the different requirements of BLER and decoding lantency in different scenes,the decoding algorithm is switched to achieve the optimal decoding performance.