In order to address the challenges posed by dense static obstacles and inefficient coordination among multiple automated guided vehicles (AGVs) in logistics sorting centres, this paper proposes a path planning scheme based on an improved A* algorithm. Firstly, the sorting environment is modelled using a grid map approach. Secondly, a mathematical model is established with the objective of minimising the total time cost (the sum of travel time and waiting time), whilst simultaneously considering path continuity, obstacle avoidance and conflicts between multiple AGVs. The conventional A* algorithm is subsequently enhanced by the incorporation of an obstacle detection mechanism during the path search process, ensuring that extended nodes avoid the obstacle grid. Furthermore, a priority rule based on ‘geometric distance × payload’ is designed to dynamically assign priorities to AGVs. When combined with real-time grid occupancy detection, the proposed method is capable of predicting and dynamically resolving conflicts between multiple AGVs, effectively handling intersection, head-on and blocking conflicts. After running the algorithm on a 20-meter-by-20-meter grid map containing 78 static obstacles and 3 AGVs, it was found that the improved algorithm could generate obstacle-free paths for all AGVs within 0.05 seconds.
The efficacy of the priority mechanism in resolving conflicts at intersections has been demonstrated, leading to a reduction in both waiting times and total system time. In comparison with conventional methods, the proposed approach enhances multi-AGV coordination in environments characterised by a high density of obstacles. The findings provide substantiation for the efficacy of this model and algorithm for practical applications in sorting centres.