Improved Conflict-Based Search for Multi-Agent Path Planning with Spatiotemporal Constraints

Multi-Agent Path Finding (MAPF) has gained increasing attention in the field of robotics and AI due to its extensive applications. However, traditional solutions often fall short in real-world scenarios, struggling with the complexity of continuous workspace and relying on simplified models. In response, ICBS-H is proposed as a novel MAPF solver that improves Conflict-Based Search (CBS) by integrating body conflict, as well as spatiotemporal and kinematic constraints. Additionally, the search process of ICBS-H is accelerated through admissible heuristics, significantly reducing search times while maintaining solution quality. Compared with two established MAPF planners in simulated environments with varying numbers of robots, it is observed that ICBS-H consistently maintains a planning success rate of over 90% across all scenarios, outperforming the other methods. Field tests at Yantai Port demonstrate the practical efficacy of ICBS-H, highlighting its potential in tackling complex, real-life logistics challenges.