Preamble-Based Noncoherent Synchronization in Molecular Communication: A Machine Learning Approach

Seok-Hwan Moon, Pankaj Singh, Sung-Yoon Jung

In the field of wireless communication, there is growing interest in molecular communication (MC), which integrates nano-, bio-, and communication technologies. Inspired by nature, MC uses molecules to transmit data, especially in environments where EM waves struggle to penetrate. In MC, signals can be distinguished based on molecular concentration, known as concentrated-encoded molecular communication (CEMC). These molecules diffuse through an MC channel and are received via ligand–receptor binding mechanisms. Synchronization in CEMC is critical for minimizing errors and enhancing communication performance. This study introduces a novel preamble-based noncoherent synchronization method, specifically designed for resource-constrained environments like nanonetworks. The method’s simple, low-complexity structure makes it suitable for nanomachines, while machine learning (ML) techniques are used to improve synchronization accuracy by adapting to the nonlinear characteristics of the channel. The proposed approach leverages ML to achieve robust performance. Simulation results demonstrate a synchronization probability of 0.8 for a transmitter-receiver distance of 1 cm, given a molecular collection time duration four times the pulse duration. These results confirm the significant benefits of integrating ML, showcasing improved synchronization probability and reduced mean square error. The findings contribute to the advancement of efficient and practical MC systems, offering insights into synchronization and error reduction in complex environments.