Abstract
Molecular communication (MC) is a promising paradigm for information exchange among nanomachines. In practical scenarios, MC transmitters are often energy-constrained and thus inherently imperfect. In diffusion-based MC systems, inter-symbol interference (ISI) is a major performance-limiting factor, and transmitter imperfections further exacerbate ISI, making reliable signal decoding more challenging. To address these issues, this paper investigates an energy-efficient transmitter design that transfers molecules from internal reservoirs into the external environment by consuming free energy. The performance of the MC system with this imperfect transmitter is comprehensively analyzed under various interference conditions. Furthermore, to mitigate ISI, a convolutional neural network (CNN)-based detection scheme is proposed, which utilizes the difference in detected molecule counts as input features. Simulation results validate the effectiveness of the proposed approach in improving detection accuracy in MC systems with imperfect transmitters.
Funding Information
  • China Postdoctoral Science Foundation (2023M732877)
  • Natural Science Basic Research Program of Shaanxi (2024JC-YBQN-0649)
  • National Natural Science Foundation of China (62501482)

This publication has 28 references indexed in Scilit: