Demonstration of Indoor Resilient Robot Communication in Beyond 5G networks: Leveraging Machine Learning and In-band Full-Duplex Relay

This paper demonstrates resilient robot communication in an indoor environment with non-line-of-sight (NLOS) communication. By using machine learning (ML), we predict the received signal strength indicator (RSSI) using data from cameras and Light Detection and Ranging (LiDAR) equipped on the robot, enabling dynamic adjustments to the robot’s path to reduce control disruptions. Furthermore, an in-band full-duplex (IBFD) relay with a self-interference cancellation (SIC) is employed to overcome signal blockages and ensure low latency. Experimental results show that our machine learning (ML) model, using a three-dimensional Residual Network (3D ResNet), achieves a root-mean-square-error (RMSE) of 2.84 dB and 74.36% accuracy in RSSI prediction. Additionally, the IBFD relay system effectively achieves self-interference suppression of approximately -90 dBm. Our approach ensures reliable, uninterrupted communication, highlighting its potential for robot control in complex environments, particularly for IMT-2030-based applications.