Autonomous mobile agents such as autonomous cars and transporters are becoming a reality which brings needs to coordinate them especially in cluttered environments such as road junctions and warehouses. There exist formal models and techniques for so-called multi-agent path finding that seem appropriate for this problem. However, these models were proposed for theoretical environments and it is not clear how they fit real robots in real environments.
The aim of this project is bridging these techniques and real needs. This will be implemented by translating theoretical plans to control procedures of robots, evaluating execution of these plans on real mobile robots, and proposing novel formal models reflecting the reality better, together with solving techniques for these models. A semi-centralized approach will be used, where the coordination of multiple agents will be implemented using centralized approaches producing robust solutions with some freedom added in translated control procedures to handle dynamicity of environment.