한국통신학회논문지
Ordered Layered Belief Propagation with Partial Updates for LDPC Decoding
Hyojeong Choi, Gangsan Kim, Hong-Yeop Song
Abstract
We propose a low-complexity LDPC decoding strategy by modifying the conventional layered belief propagation scheduling. The proposed strategy prioritizes check nodes according to their residual sums and updates only a subset with larger values in each iteration. We call this approach ordered layered belief propagation (OLBP) with partial updates. This method effectively reduces the number of message updates per iteration compared with conventional schemes, while maintaining comparable error-correction performance. Simulations on the GPS L1C LDPC code using the Sum-Product (SP), Normalized Min-Sum (NMS), and Offset Min-Sum (OMS) algorithms show that OLBP achieves nearly identical FER performance with only 70% of the message updates at 10 iterations, and 50% at 20 iterations, demonstrating substantial complexity reduction without noticeable performance loss.