Self-healing Algorithm Design And Implementation For Mobile Robot Formation

Compared to the traditional single mobile robot system, the multi-robot system, as a redundant system, has reconfiguration ability and structure flexibility, thus can reduce the cost and improve the efficiency and robustness of the system.Multi-robot formation is the technology to control multiple autonomous mobile robots to maintain a specific shape during the movement. It is widely used in military reconnaissance, disaster relief, aerospace and industrial applications, etc. In military applications, multi-robot formation can be used for many tasks such as area coverage, breaching defences and so on; In disaster relief, the robot formation can enlarge working scope and improve the efficiency of the rescue; In the field of aerospace, the small satellite formation can save the fuel for propulsion and expand the sense capabilities of the satellite cluster; In the intelligent transportation, the formation of driving vehicles can greatly alleviate the traffic of the highway; In the biological field, formation control could help researchers understand social behaviors of livings, such as the collective migration phenomenon of birds or fishes.However, it is inevitable that part of robots in the multi-robot formation may fail due to the dangerous or hostile working environment with a lot of uncertainty. The failure of robot will degrade the performance of the formation, reduce the working efficiency, even lead to the breakdown of the mission if without corresponding measures. Therefore, it is urgent to design a self-healing algorithm that can autonomously repair the topology of the formation and improve the performance of the system when robots fail in the formation.In this paper, a gradient based self-healing algorithm is proposed for the problem of self-healing robot formation with failed robots. The proposed algorithm enables robots to autonomously repair the formation topology and improve the synchronization of the network under local interactions among neighboring robots. The contributions in this paper are as follows:1. Considering the effect of different robots on the motion synchronization of formation, a gradient generation and diffusion mechanism is proposed to form the gradient distribution in the robot formation; then, combining the proposed gradient generation and diffusion mechanism and the switched topology theory, self-healing rules and the corresponding distributed individual control algorithm are presented to repair the network topology and improve the performance of motion synchronization; furthermore, the improvement of the proposed method on the motion synchronization and the optimality of the self-healing path are proved through theoretical analysis;2. A gradient vector based recursive self-healing algorithm framework is proposed to solve the problem of repair robot sequences optimization under different indices. The optimization of different indices can be achieved by changing the content of the gradient vector which is derived from the concept of gradient value.3. The effectiveness and optimality of the proposed method is validated by numerical simulations under different network topology models with a number of robots and by the self-healing formation experiment in the real multi-robot platform.