Research On Path Planning And Path Tracking Based On Four-wheel Vehicle

With the rapid develop of mobile robot technologies in the past ten years; Mobile robot is playing a more and more important role in human's life. Mobile robot can replace humans to do some dangerous works, especially in these places where humans cannot approach. High intelligence of a mobile robot is required for it to handle problems in the complex and severe environment. The automatic driving of a mobile robot is the base of its higher level tasks, ensuring the mobile robot reaching the designated target point.When driving in a complex environment, the mobile robot must follow a path which is wide and smooth to drive safely. In order to fulfill this goal, the two methods of Voronoi graph and support vector machine are combined together, then a new method for path planning based on support vector machine is proposed in this paper. As a path planning method, Voronoi graph has the advantage of fast computation speed, and the path created by it has the merits of highly safe. Because of these two advantages, the Voronoi graph is employed as a preprocessor to roughly find a path connecting the initial and goal in this paper. However, the path generated by the Voronoi diagram is composed of straight line segments, the path don't satisfy the need of the smoothness, so the support vector machine of with Gauss kernel is used to smooth the path. Due to advantages of the SVM with RBF kernel, the post-processed path has the merits of both smoothness which is ensured by Gaussian kernel, and wide clearance which is ensured by the decision function of SVM. First, the obstacles are treated as particles, and the mechanism of the algorithm is described.By giving different values of the parameters of the RBF kernel,comparison of the advantages and disadvantages between the generated paths is made to study the effect of the RBF kernel on the formation path.It is proved that only the proper selection of the parameters of RBF kernel can the algorithm generate a satisfactory path.In order to verify the effectiveness of the algorithm, the simulation is also carried out in the environment with polygon obstacles.The effectiveness and applicability of the path planning method are verified by simulation in different environments.Another important task in the automatic driving of the mobile robot is the path tracking, which ensures that the moving robot follow the safe planning path. In this paper, a four driving vehicle is studied and its kinematics model is made.Because the mobile robot is anonlinear and non-fully constrained system, some classical control algorithms cannot have a good control of it. The fuzzy control method is adopted in this paper.A fuzzy controller is designed for path following of the mobile vehicle, which is composed of a steering fuzzy controller and a speed fuzzy controller.The steering fuzzy controller has the inputs of the horizontal offset error and the heading angle error,and the angel generated by it is provided to control the bottom agent.The speed fuzzy controller ensures the speed in a safe range depend on the steering angle and the curvature of the planning path, which ensure the safety of forward speed.In this paper, the simulation system of the fuzzy controller is established in the Matlab environment,and the performance of each parameter in the system is analyzed and the effectiveness and applicability of the fuzzy control algorithm are verified by following the given path.