Adaptive Successive Cancellation List Decoding for Polar-Coded Probabilistic Amplitude Shaping

The multi-level coding-based Polar-coded Probabilistic Amplitude Shaping (PAS) involves a pre-coding step for systematic encoding. Due to this, the Cyclic Redundancy Check (CRC) at the receiver is available only post-decoding of all the bit-levels; this delays the CRC-based adaptive list-size decision. Therefore, the presented approach utilizes PAS symbol distribution for individual bit-level path-checks that can determine the insufficiency of a list-size at an early stage. Also, our system model involves the necessary Polar-coded PAS design modifications to enable the Block partition labeling. We compare the average decoding complexity and the Frame error-rate performance with other real-time list-size decision approaches. Furthermore, we show that in comparison with the adaptive CRC-aided successive cancellation list decoding, our proposed algorithm shows 27% less average latency while achieving better error correction performance for short block-lengths.