Interference Level Estimation Using Machine Learning in a Robotic Force-Seeking Scenario
Richard Candell, Mohamed T. Hany, Karl R. Montgomery, Yongkang Liu, Sebti Foufou
Wireless communications plays an essential role in the future cyber-physical systems vision which includes having more sensors and actuators, and, hence, more information transferred through wireless. In this article, we consider an industrial use case of a robot arm control system equipped with a force-torque sensor. Movement of the arm is tracked by a vision-based ground truth measurement system. Movement of the arm is controlled by a robot controller applying a downward pressure on a spring assembly until a predetermined force is detected. The remote vision-based observer provides readings about the position of the robot arm where these readings are used to estimate the signal-to-interference ratio of the wireless link. A supervised machine learning approach is used for the wireless channel quality estimation. In this paper, we study the impact of various system features on the performance of various machine learning algorithms and compare their performance. Moreover, we investigate the impact of the training and the estimation period on the performance of the proposed approach. The results provide insights about the impact of wireless communications on cyber-physical systems and an example of employing machine learning to improve industrial wireless deployments.