Channel Modeling for Mobile Molecular Communication With Anomalous Diffusion by Deep Neural Network

Area of Science:

• Electrical Engineering
• Computer Science
• Biomedical Engineering

Background:

• Mobile molecular communication (MMC) systems show promise for nanoscale applications, especially with anomalous diffusion.
• Accurate modeling of anomalous diffusion channels with multiple receivers in MMC is challenging.
• Existing analytical methods for channel impulse response (CIR) are limited to normal diffusion and static conditions, failing to adapt to complex, time-varying environments.

Purpose of the Study:

• To develop and evaluate a novel deep neural network (DNN) based method for predicting the CIR parameters in a 3D MMC system with multiple receivers under anomalous diffusion.
• To compare the performance of the proposed DNN model against Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) models for CIR prediction in anomalous diffusion scenarios.

Main Methods:

• A three-dimensional (3D) mobile molecular communication (MMC) system model with one transmitter and multiple receivers was established.
• A deep neural network (DNN) was designed and trained to predict the channel impulse response (CIR) parameters under anomalous diffusion conditions.
• The predictive performance of the DNN model was compared with RNN and LSTM models through simulations.

Main Results:

• The DNN-based model significantly outperformed both RNN and LSTM models in predicting the CIR parameters across various anomalous diffusion conditions.
• The proposed DNN approach demonstrated robust and accurate prediction capabilities for the complex channel characteristics of MMC systems.
• Simulation results validated the effectiveness of the DNN model in improving CIR prediction accuracy.

Conclusions:

• Deep neural networks offer a powerful and effective approach for modeling anomalous diffusion channels in mobile molecular communication (MMC) systems.
• The developed DNN model provides a significant advancement in accurately predicting CIR, crucial for reliable communication in complex nanoscale environments.
• This work presents a novel methodology for enhancing channel modeling in MMC systems, particularly those operating under anomalous diffusion conditions.