Synthesizing LDPC Belief Propagation Decoding with Molecular Reactions

This paper proposes a CRN-based implementation approach for low-density parity-check (LDPC) decoding based on belief propagation (BP). Since the belief (probability) can be naturally mapped to molecule concentration, LDPC decoding can be realized with CRNs instead of silicon based hardware. Theoretical analysis and numerical simulations have demonstrated the feasibility of the proposed approach. Note that, we do not try to substitute the silicon-based LDPC decoder with CRN- based one for high-speed applications. We show that this method can be generalized for other BP-based algorithms and is suitable for large-scale, bio- interface, and latency-insensitive applications.