QL-Based Adaptive Transceivers for the IoBNT Communications

This paper introduces an adaptive transceiver scheme for bio-nano things (NTs) situated within blood vessels communicating through a time-varying molecular channel. The proposed scheme employs a Q-learning-based adaptive transceiver (a so-called QL-ADT), wherein an agent gradually learns how to adapt the transmission parameters to the current state of the channel. A real heart rate dataset is used to estimate the blood flow velocities over time, based on which a time-varying molecular channel is modelled. In the practical implementation of the QL-ADT, an external gateway, situated on the skin, monitors the body’s heart rate over time and interfaces with the NTs through implantable nano devices. The gateway dynamically adjusts the communication parameters of the NTs based on the measured heart rate and what it has learned during the training phase. The proposed QL-ADT scheme showed significant improvement in the achievable raw bit rate (RBR) and error performance for a real heart rate dataset.