Digital communication free collaborative robots swarms – RoboSCo

This project presented in response to the ASTRID ASRO 2021 call for projects deals with perception and decision mechanisms for robot swarms with neither absolute positioning nor digital communications. It aims at studying :

- Perception mechanisms and algorithms allowing a swarm of autonomous robots to locate and understand a dense and rapidly changing environment without the use of an external positioning system, with limited on-board processing resources within each robot.

- Mechanisms and algorithms for making individual strategic decisions in order to achieve a coherent collective strategy in an environment prohibiting digital communications and changing rapidly.

Both these scientific objectives have a common goal : getting closer to human perception and decision-making mechanisms in order to increase robustness of a robots swarm to digital communication interferences and to allow interaction with humans. For this reason, by mimicry with humans, project introduces a virtuous constraint by prohibiting the use of digital communications, between robots in all their forms (radio, LIFI, ultra-sonic ... ) and as Global Navigation Satellite Systems (GNSS).

This lack of digital communication does not mean a global lack of communication. As for humans, positioning on a field and performing actions can be used by other robots as meaningful information on the behavior of the considered robot. This is a mode of parsimonious communication allowing collaboration among the swarm of robots without radio link, and offering prospects for interactions with humans. Many applications are concerned : military but also civilians such as in hospitals where robots from different manufacturers and humans will collaborate in a near future.

This collaboration without digital communication requires increased abilities of each robot for perceiving its environment to overcome the lack of GNSS. First main research axis of the project will be finding innovative artificial intelligence algorithms allowing a robot to locate itself in its environment and understand the state of a field of operation and its evolution, with only limited on board computing resources and without GNSS.

The lack of digital communication also leads each robot in the swarm to receive asymmetric, partial and uncertain perception. The search for individual decision-making algorithms in order to contribute to a coherent group strategy without having digital communication in a permanently evolving context will be the second main research axis of this project.

Both these researches will be jointly tested and validated during the project using simulation tools and in a real conditions in the context of autonomous soccer games using RoboCup type robots (Middle Size League category). This project will take advantage of experience and achievements of the robotics team of Toulon University, led by IM2NP and COSMER laboratories. This team is the only French team qualified and participating in RoboCup, the robotics World Cup, in Middle Size League category.