Numerical and Experimental Assessment of Distributed Model Predictive Control of Multi-agent Circular Formation

This paper presents the development of a didactic robotic demonstrator to study the circular formation problem (CFP) of multi-agent systems based on Distributed Model Predictive Control (DMPC). CFP can arise in various fields, such as robotics, drones, and space applications. This study proposes a method for multiple agents to sequentially enter a circular orbit, one by one, and move along the orbit at a predefined speed while maintaining a desirable distance from each other. Due to the advantages of DMPC, the proposed method allows each agent to independently generate control inputs while considering various types of constraints related to collision avoidance with the other agents. The methodology is tested and validated with three differential drive robots, i.e., Turtlebots.