A Probabilistic Constellation Shaping Scheme for Molecular Communications

Molecular communication (MC) conveys information via chemical signals between nanomachines, which lays the foundation for the Internet of Bio-Nano Things. Based on the fact that probabilistic constellation shaping (PCS) enables reliable communication by improving energy efficiency, in this paper, we propose molecular shell mapping (MSM) to implement PCS for MC systems. The MSM method employs the concentration se-quences with the lowest sum sequence weight to construct energy-efficient signal constellations, and designs a sorting principle for sequences with the same sequence weight to mitigate inter-symbol interference (ISI). We design a Genie-aided maximum-likelihood (ML) detector and a realistic ML detector to leverage the constructive effect of ISI from the current concentration sequence, and derive their achievable rates. Numerical simulation results confirm that the shaped signaling using the MSM method is more energy-efficient than the equiprobable signaling, achieving an up to 1.5 dB gain at an ultra-short blocklength.