A machine learning-based concentration-encoded molecular communication system

Abstract

Molecular communication (MC) is a recent novel communication paradigm, which could enable revolutionary applications in the fields of medicine, military, and environment. Inspired by nature, MC uses molecules as information carriers to transmit and receive data. Concentration-encoded molecular communication (CEMC) is an information encoding approach, where the information is encoded in the concentration of the transmitted molecules. In this paper, we propose a machine learning (ML)-based CEMC system. In particular, we propose a modulation scheme named concentration position-shift keying (CPSK), which encodes information as the position of the transmitted molecular concentration. After passing through a diffusion-based channel, the molecules are captured via a ligand–receptor binding process (LRBP) at the nanoreceiver. Then, a ML-based approach is employed to decode the data bits. From numerical simulations, it has been shown that increasing the transmission time and using 4-ary CPSK would enhance the communication performance of the proposed ML-based CEMC system. In addition, we found that the ML receiver mitigates the bias effect and reduces inter-symbol interference (ISI) of the diffusion-based molecular channel. As a result, the proposed ML-based receiver shows better performance than the conventional maximum-likelihood (MLE) receiver.