Xingyu Chen

This paper investigates an explicit polar coding scheme for a two-user discrete memoryless multiple access wiretap channel with partial rate-limited feedback (MAC-WT-PLF). Feedback is exploited to increase channel-input correlation, inject dummy messages, and encrypt messages using a one-time pad. Existing MAC-WT polar coding schemes assume independent channel inputs, which cannot support the correlation requirement. Therefore, we propose an explicit polar mapping that leverages feedback to introduce correlation between the channel inputs. This mapping allows both the transmitter and the receiver to restructure the correlation in opposite decoding directions. The proposed scheme relies on source polarization, block Markov coding, superposition coding, lossy source coding, and ratesplitting, without making symmetry or degradation assumptions on the channel model. Rigorous information-theoretic asymptotic analysis establishes that the proposed scheme ensures both reliability and strong secrecy, and attains the entire achievable secrecy rate region given in prior work.

Channel state information (CSI) is differently available at each terminal in state-dependent wiretap channels (SD-WTCs). Considering a random channel state non-causally available only at the encoder, this paper investigates an explicit polar coding scheme for the discrete and memoryless SD-WTC, where Alice aims to transmit a secret message (SM) and a secret key (SK) to Bob while concealing them from Eve. Based on a two-layer superposition coding, which includes an inner layer and an outer layer, the proposed polar coding scheme can achieve the inner bound of the current optimal SM-SK capacity region for the discrete and memoryless SD-WTC with CSI non-causally available only at the encoder. A cross-layer construction is proposed to address the issue where the regular chaining construction used in wiretap polar codes cannot be applied in the inner layer. Results show that the decoding error probability (DEP) of the legitimate decoder vanishes to zero as the block length increases. In addition, the proposed scheme is proven to satisfy strong secrecy.