Optimization of Iterative Decoding for Polar Product Codes

To improve the error-correction performance of polar product codes, this paper designs an optimization criterion for the weighted factors (WFs) in the iterative decoding. Firstly, a weighted mean square error criterion is proposed to optimize the WFs. Secondly, this criterion for optimizing the WFs is extended to polar and single parity-check turbo product code (P&SPC-TPC) and polar turbo product code (Polar-TPC). The optimized WFs obtained according to this criterion are superior to those obtained from Monte Carlo simulations. The two optimized iterative decoders of TPCs, P&SPC-TPC and Polar-TPC, can obtain noticeable performance gains compared with the original product decoders.