Research On Path Planning For Multiple Mobile Robots In Intelligent Warehouse

In the field of electronic commerce, intelligent warehouse using multiple mobile robots will gradually become a development trend of the logistics industry. Multi-robots path planning technology is the key technology of intelligent warehouse system. With the increasing number of warehouse area and robots, its importance is obvious. The thesis research on it, the purpose is to design a suitable for intelligent warehouse environment of multi-robots path planning algorithm.This paper makes a research on global path planning and local path planning for intelligent warehouse multi-robots. The robots system uses hybrid control architecture for multi-robots, and the central controlling unit is responsible for the global path planning. The algorithm of global path planning is an improvement on traditional genetic algorithm, which uses link diagram method to establish the warehouse environment model. Local search is used to form initial population in the algorithm, and then new chromosomes are created by crossover and mutations. The fitness function matrix is established in it, which is fully considered the path length and the time needed for mutual cooperation between multi-robots. Through the comparison of chromosome fitness function size, the global optimum path is searched.Local path planning for multi-robots through communication negotiate is made by the decision module robot itself, where algorithm is an improvement of SHAA neural network algorithm model. Dynamic mathematical model makes use of the characteristic of step function that can simplify environment. And introduce step function into the state equations of static mathematical model of SHAA algorithm. In order to make the algorithm suitable for dynamic path-planning, each grid will be seen as neurons, using the distribution of neuron dynamic activation values to describe the complicated dynamic environment. To avoid local path conflict among multi-robots, the2D grid map is extended to3D space time map. And then the reservation form is introduced for local path planning of robots. Because of the difference between controllers for global and local path-planning, maps of environment model, map transfer function is used. The global map refines to the local environment map.In this paper, simulation system is established to build the environment of intelligent warehouse work and to simulate the motion state of the robots. At last, the proposed algorithm which is proved by simulation system is suitable for path planning of multiple mobile robots in intelligent warehouse.