Research On Global Path Planning And Trajectory Tracking For Mobile Robot

The research of mobile robot rapidly develops towards autonomization andintelligence with the development of computer science and sensor technology. Mapacquisition is the basis of the target point exploration in the environment. The robotwhich receives a task to find a collision-free path in obstacles must have anautonomous navigation capacity. Therefore, global path planning of mobile robot isalways been one of the hottest spots of robotics. The omnidirectional vehicle has beingused by an increasing number of researchers due to its advantages in operability andnavigation distance. The main interest of the present work is concentrated on pathplanning issues of omnidirectional vehicle.The path planning issue of the robot can be divided into three parts:environment modeling, path planning and trajectory tracking. The Kinect sensor isused to obtain environmental information because it is economical and applicable.Then, environmental map is built by the usage of a clustering algorithm DBSCAN anda fitting method Minimum-Perimeter Polygon. Before the path planning begins, theachieved environmental information should be converted into some map informationthat the robot can read. After that, the free space method is used to build the motionenvironment of the robot. In this paper, a path planning algorithm is presented basedon A*algorithm and Bezier curve to create a smooth and collision-free path. Based onthe dynamic and kinetic characteristics of the omnidirectional vehicle, the motiontrajectory of the robot is controlled by Bang-Bang control strategy. In the final part, thepresented algorithm is developed in simulation and reality condition. The proposedscheme can improve efficiency in robot navigation as indicated by the experiments.