Robot Motion Planning Based On Rapid-Exploring Random Trees In Complex Environment

Robotics is the research of the technology associated with the theory, design,manufacture and application of intelligent machines. With developments in this technology,the application of robotics is becoming more and more extensive; and mobile robot is onetype of robot that is rapidly increasing in relevance, application and demand. Effective pathplanning techniques are required to develop navigation algorithms for successful mobility ofrobots in their particular environments. The Rapidly-Exploring Random Tree (RRT) techniqueis a search system used in path planning. This system can generate new nodes in order toreach the goal point. It makes the algorithm easily meet the needs of the kinetics anddynamics constraints of the system and it shows a good performance in tackling the pathplanning problems in recent years.The basic RRT algorithm is modified and a new algorithm is proposed called Bias-goalRRT algorithm. The two algorithms are compared through computer simulation. Results showthat the modified one has taken care of some of the shortcomings of the basic one such aslong searching time and non-heuristic searching and it can increase the speed and efficiencycompared with the basic one.The modified RRT algorithm is applied to solve two kinds of path planning problems incomplex environment. Both of the two problems consider the kinetic model of the wheeledrobot. The former one uses the modified RRT algorithm mainly to solve the path planningproblems in dynamic environment of the mobile robot with nonholonomic constraints. Thelater one uses the algorithm combined with modified RRT and artificial potential field toavoid the shortcomings of the artificial potential field method, such as local minimum. Thecombined algorithm uses artificial potential field for the local planner while the RRTalgorithm is used for global path planning. With the help of the two algorithms, local minimaare avoided. The simulation results show that the above two methods are valid and can beapplied in path planning for wheeled mobile robots.