Collision Avoidance Pathfinding of Multiple AGVs Considering Motion Uncertainties

The multiple automatic guided vehicles (AGVs) pathfinding methods that do not consider the uncertainties of the future motion state may lead to large-scale congestion. To address this problem, this paper proposes an improved conflict-based search (CBS) algorithm with a load-information map. Firstly, the grid map with load information and the evaluation criteria of collision-free pathfinding is constructed. Then, an improved CBS algorithm is designed under a set of constraints on the motion states. This is achieved by performing a search on the constrained conflict tree between individual robots using load-information-map at the high level while exploring efficient single-robot searches with modified penalty function at the low level. Finally, the proposed improved CBS algorithm is tested through simulation and experiment. The results show that it not only finds a collision-free path for multiple AGVs under the motion uncertainties but also obtains a comprehensive optimal solution that considers the evaluation criteria and map load simultaneously.