Research On Robot Cars' Path Planning Algorithms In Warehouse

The development of electrical business booms the automation/intelligent level of Logistics. Since AMAZON used KIVA system to improve the efficiency of logistics process, domestic and foreign companies have begun to research intelligent logistics system,and most of them are similar to KIVA system. Taking KIVA system's warehouse layout as an example, this paper aiming to solve the problem of multiple robots' path planning have the following main contents.First, this paper solves the multiple robots path planning problem with three different algorithms: According to the characteristics of KIVA system's warehouse layout, this paper improves the traditional artificial swarm fish algorithm with bring in the fish's leave-group strategy which can reduce the amount of calculation of artificial swarm algorithm.Combining with the collision avoidance rules, the paper shows a feasible way to plan multiple robots' path based on the improved artificial swarm fish algorithms. To reduce the number of invalid path based on GA, this paper divide the whole path into N parts, so the robot only have to plan the following few steps. Combining with the collision avoidance rules,the paper shows a feasible way to plan multiple robots' path based on GA.To improve the traditional A* algorithm, the paper redefined the close list through bringing in the past list and frequency list. Combining with the collision avoidance rules, the paper shows a feasible way to plan multiple robots' path based on improved A* algorithm.Second, to illustrate the three algorithms' feasibility, the paper simulates the multiple robots' path planning process and the scale of the warehouse the robots working in is about 150000 square meters which is as large as the Amazon's real warehouse .The result shows that GA's path planning effect is the worst. The average path planning time for one step of improved artificial swarm fish algorithms is shorter than the other two algorithms'. As to the time and the total length for all the robots to finish all tasks, the improved A* algorithms' is shorter than the other two algorithms'.This paper has practical significance to improve the efficiency of logistics operations, and a certain reference value to the domestic enterprises' research on the use of intelligent logistics system.