Research On Path Planning Method Of RoboCup Middle Size Soccer Robot

Robot soccer game as an important branch of artificial intelligence research, through the active participation of many experts and scholars in recent years has been considerable development and training a large number of related areas highly qualified experts and access to a large number of research results and promote the artificial research advances in intelligent,a lot of research in order to put into practice, the harvest of fruitful results. The path planning as an important research branch of robot soccer game has an extremely important role in the whole research system, and have highly value in the actual production and life, so this paper choice path planning as the research object.In this paper, RoboCup middle size group robot soccer competition for the object,studied the problem of robot path planning method. This article first football game with robots used MT-OR robots were introduced, and then for the application objects and application environment were analyzed, using the rapidly exploring random tree algorithm as the global path planning algorithm and using the artificial potential field algorithm as the partial path planning algorithm, summed up both the advantages and disadvantages of them after simulation and analysis. For the problem that rapidly exploring random tree algorithm for fast traversal requires global random search with a poor real-time, drawing lessons from artificial potential field method virtual a gravity from the target point to other point to attract the random tree growth tend to the target point which greatly improved the algorithm real-time, and set a simple path smoothing algorithm; for the problem that artificial potential field is not suitable for dynamic environments, set a path planning algorithm based on improved dynamic potential field algorithm, adding the speed repulsion potential function and acceleration repulsion potential function to original repulsion potential function and proved the improved algorithm could achieve path planning in dynamic environments after simulation and analysis.Finally, set a new algorithm through combine two improved algorithms,solve the disadvantage of using artificial potential field algorithm to path planning that the planning path always far from the optimal path and the local minimum point problemand the goal impossible reach problem and the problem that couldn't meet the real-time request when using rapidly exploring random tree algorithm alone. Simulation on the MATLAB as this condition that our player moved to the target point with the intercepting of enemy players simulation.Simulation results show that the mixed algorithm based on improved rapidly exploring random tree algorithm and improved artificial potential field algorithm can successfully path planning and the planning path is more excellent and could meet the real-time request of robot soccer competition.