Path Planning And Tracking Control For Wheeled Mobile Robot

In some specific environment the mobile robot can assist people to do some relatively simple mission, such as the office service robot, the museum tour guide robot and so on. Some mobile robots can also work in the dangerous and complex environment to finish the missions that cannot be done by human beings, such as the explosive-handling robots, search and rescue robots and so on. Among all of the mobile robots, wheeled mobile robots have simple structure and flexible operation ability, and have wide application.Being able to detect and obtain a safe path from the starting position to the target position is the prerequisite for the autonomous wheeled mobile robot to finish its missions. General speaking, the requirements of the path for the wheeled mobile robot are that by running in the planned path the mobile robot can get to the target position avoiding any obstacles. At the same time the length of the path should be the shortest. Move over, the wheeled mobile robot has some inertia. If the path that the wheeled mobile robot runs has much turning, the efficiency of the mobile robot can be largely decreased, more energy is used to run and control the mobile robot, and sometime the performing stability of the wheeled mobile robot can even be influenced. For this reason, the proposed path planning algorithm in this thesis should get the optimal path with less turning. In this thesis the modified ant colony optimization algorithm is employed to do the path planning for the mobile robot. Ant colony algorithm has the advantages of positive feedback, distributed computing, and self-organization, and is widely used in the wheeled mobile robot path planning problems. But ant colony algorithm has also some problems, such as easy to fall into local optimal solution, stagnation and slow convergence speed. Based on the advantages and problems of the ant colony algorithm and the special requirements of the path planning for the wheeled mobile robot, we propose a modified ant colony algorithm. In order to get short and less turning optimal path, when calculating the heuristic function not only the distance factor but also the mobile robot turning factor are considered, which makes the obtained optimal path is both short and with less turning angles. The modified ant colony algorithm can finish the path planning task for the mobile robot well.After obtaining the planned optimal path, in order to finish the mission well, an accurate and robust path tracking control system is very important for the wheeled mobile robot. In this thesis the neural network is used to estimate the model of the wheeled mobile robot. Based on the neural network model the path tracking control algorithm for the wheeled mobile robot is designed. Neural network can estimate the nonlinear system very well. At the same time, the neural network has good adaptability and robustness by online training. In this paper the neural network based controller is designed by using the kinematic model of the wheeled mobile robot. Simulation is used to verify the designed controller. In the simulation various paths is considered. The good simulation results indicate that the proposed neural network controller can provide good control results for mobile robot.