Graph-based coordination algorithm for robotic mobile ad-hoc network in collapsed mines

One of the main problem arising in case of teleoperated robots moving inside a collapsed mine in order to rescue survivors is that a remote operator cannot usually rely on affordable communication systems. Having a robust communication network is the key point to let an external rescue team operate without any risk, thus improving the chances to find survivors. The solution proposed in our paper is based on the coordination of a mobile ad-hoc network. In particular, we have considered a network of mobile antennas (i.e. small size mobile robots equipped with a wi-fi router and proximity sensors) controlled in order to preserve robust and reliable wireless communications to broadcast control inputs from the operator to the remotely connected robot and, at the same time, to gather environmental data. Based on concepts borrowed from graph theory, our new control algorithm ensures that mobile antennas are able to autonomously optimize their spatial distribution inside the mine in order to guarantee transmission redundancy and fault tolerant properties for the communication network. At the top of the control hierarchy used to coordinate robotic antennas, an algorithm base on virtual nodes has been implemented in order to avoid collisions among robots and between the robots and environmental obstacles. This approach to obstacle avoidance has been introduced because it can be included in the same graph-based framework used to coordinate antenna robots. Moreover, thanks to the communication redundancy, a significant amount of data can be transmitted and thus even the position of obstacles detected by one robot can be broadcasted to teammates. The proposed approach has been validated by means of mathematical demonstrations and numerical simulations on Matlab/Simulink. Future developments will consider real experiments.