Connecting Robots with Concurrent Exploration of Control and Communications

Abstract

Multi-robot systems (MRS) have many applications and the efficient operation of MRS relies on coordination of robots. However, it is difficult to build network connections among randomly distributed robots in the presence of robot movements and weak wireless channels. In this work, we propose to jointly exploit communications and motion control to efficiently establish robot connections. To achieve this goal, we concurrently use MUSIC and particle filter to more accurately and efficiently estimate robot signal directions, built on which signal strength-based potential field is formed to control robot motion to establish and maintain communication links. Our studies based on test bed and simulations demonstrate the effectiveness of our algorithm in networking robots, with much higher number of robots connected compared to peer algorithms.