Machine Learning-Based Detection Time Estimation for Molecular Communication

Molecular communication (MC) utilizes the release, diffusion and reception of molecules to transmit information. It has promising prospects in the field of drug delivery. The detection time estimation of the receiver in MC system plays important roles in the resource consumption at the receiver. Existing strategies of traditional detection time mainly focus on known channel state information (CSI). In this paper, we propose a method for estimating the detection time of the receiver in MC system with unknown CSI by using deep neural network (DNN) model. We employ the Monte Carlo simulation to capture the positions of molecules in the three-dimensional environment. The dataset is generated based on the coordinates of the molecules at each position. The numerical results show that the detection time can be accurately estimated by the DNN model which exhibits good detection and generalization abilities. In addition, the number of molecules released by the transmitter and the minimum distance between the transmitter and the boundary of the receiver have impacts on the accuracy of detection time estimation of the receiver.