An Error-Threshold Based Joint Phase Estimation and Decoding Method for Non-Pilot-Aided Systems

In 5G/6G communication environments, synchronization technologies encounter challenges including excessive pilot overhead, poor reliability at low signal-to-noise ratio (SNR), and limited estimation accuracy. This paper proposes a joint carrier recovery and iterative decoding algorithm based on weighted and normalized syndrome satisfaction probability (WNSSP). Our architecture implements parallel phase compensation through a multi-layer factor graph model. Within this framework, WNSSP dynamically weights and fuses belief provided by the soft low-density parity-check (LDPC) codes decoder to construct a maximum likelihood estimation function that simultaneously achieves carrier recovery and information reconstruction. To address the single-metric evaluation limitation, code-modified mean square error (CM-MSE) is established according to the residual phase error threshold, which enables concurrent assessment of synchronization accuracy and demodulation reliability. Simulations demonstrate that while effectively eliminating redundancy caused by pilot sequence, the proposed algorithm achieves 0.5-0.8 dB phase estimation gains and 0.15-0.5 dB bit error rate (BER) improvements compared with existing methods.