The problem of AGV's path planning is a very important part of the control of AGV, and a basic part of the navigation of AGV. The paper mainly focuses on the study of the global path planning in the static known environment, and the study of the local path planning in the dynamic unknown environment.The first part analyses the space-location of the AGV. A laser scanning location mechanism and algorithm are presented to locate AGV's position in the paper. Because the triangulation algorithm is difficult to implement path planning when AGV moves around, the extended Kalman Filter is used in the localization compensation.The second part proposes global path planning of AGV in the static known environment. The paper uses an improved method of the visibility graph to establish environment model, also a method of AGV global path planning with Genetic algorithm is presented. At last, AGV global path planning simulation is carried out in the environment with static known obstacle by matlab, and the result indicates that the method is feasible.The final part proposes the local path planning of AGV under the dynamic unknown environment. In the dynamic environment, the problem of AGV dynamic path planning is difficult to solve. The paper proposes a mathematical model of dynamic environment based on velocity obstacle and the concept of risk degree of collision, also a method of AGV path planning with fuzzy neural network is presented. At last, AGV dynamic path planning simulation is carried out in the environment with immobile and mobile obstacle, the simulation result indicates that the method is feasible. |