Dynamic Distributed Coordination Schemes for Multi-Mobile Robot Systems Under Collision Avoidance Constraints

Guaranteeing collision avoidance is of paramount importance in view of accomplishing missions where many agents are involved to share the same space. The main objective of this work is to expand the Turn-Based Command Governor approach in (Tedesco et al., 2019) by taking non-convex Collision Avoidance constraints into account when performing Plug-and-Play (PnP) operations among agents operating in a 2D environment. To deal with such a scenario, formal conditions that guarantee collision-free Plug-and-Play (PnP) operations are given. These conditions are based on the concept of safe areas, which define regions where agents can safely perform PnP operations without the risk of collision. The effectiveness of the proposed strategy is illustrated through various examples, highlighting its potential for mission accomplishment involving multiple agents. Note to Practitioners - This paper presents a novel distributed control strategy designed to coordinate multiple mobile agents safely and efficiently. By incorporating advanced control techniques such as Model Predictive Control and Command Governors, the proposed approach addresses the challenges of collision avoidance and dynamic formation changes. The methodology offers several potential benefits for practical applications, including enhanced safety, increased flexibility through Plug-and-Play capabilities, and reduced computational demands compared to centralized alternatives. Although implementation may require specialized expertise in control engineering, the potential impact of this research on robotics, autonomous vehicles, and other multi-agent systems is significant. This work provides a foundation for future research and development in distributed control and coordination.