Research On Intelligent AGV System Design And Path Planning Algorithm

This paper,intelligent AGV as the research platform, focuses on the research of path planning algorithm. Before path planning algorithm research, the intelligent AGV system model and location method are analyzed firstly.In the section of construction of system model, the coordinate system model, environment map model, intelligent AGV kinematics model and laser scanning rangefinder observation model are set up, to provide the necessary theoretical basis for the follow-up location method and path planning algorithm.In the section of location method, combined location method which is the combination of the relative location method based on photoelectric encoder and absolute location method based on laser scanning rangefinder is used, and the extended Kalman filter algorithm mix the data of photoelectric encoder and laser scanning rangefinder together.In the section of path planning algorithm, aiming at the shortcomings of the standard A* algorithm in the application of intelligent AGV, this paper proposes an improved A* algorithm which based on dynamic obstacles, reward and punishment mechanism, linear-curve-linear model, and the feasibility of the algorithm is verified in the matlab. The standard A* algorithm make the following improvements: according to the standard A* algorithm can only in the condition of global environmental information given to plan out a path from starting point to ending point, which based on dynamic obstacles of the A* path replanning is put forward; according to the standard A* algorithm does not consider the intelligent AGV width information, which based on rewards and punishment mechanism of child nodes generated strategy is put forward, give priority to the child node in the original path direction, and do not generate the child node of the current node in the diagonal direction adjacent to the obstacle node; according to the standard A* algorithm to generate the path inflection point, large cumulative turning angle and the path is not smooth, on the basis of the optimization of the key points, the linear-curve-linear moder is put forward, and the inflection points are replaced by the curve. The improved A* algorithm has greatly reduced the number of inflection points, the path is more smooth, which can better meet the operation characteristics of intelligent AGV.Finally, based on the research of location method and path planning algorithm, hardware and software system of the intelligent AGV are design, and hardware and software anti-interference method are put forward in view of the characteristics of the intelligent AGV working environment.