Path And Behavior Planning For Intelligent Robot Soccer

In recent years, intelligent robot soccer has become one of the hot research topics in the field of artificial intelligence and intelligent control. The robot soccer system is an environment with real-time dynamic and confrontational high complexity. It involves robotics, mechatronics, communication and computer technology, robot vision and sensor fusion, intelligent control and other high-tech, so it has attracted many experts and scholars to study at home and abroad. Robot Soccer is not only the display platform of these researches, but also the combat platform of the high-tech. The behavior and path planning is the significant part of the research of intelligent robot soccer system. To get good grades, there is not only need a good behavior planning, but also a good path planning. Since combined the two parts, the purpose of more accurate, more rapid completion of attack and shooting can be achieved.This paper focuses on the research of robot shot behavior and path planning, and it aimed at resolving the shortcomings of inadequate use of the shot data, improving the effectiveness of algorithms of the path planning. The improvement of the shooting behavior mainly focuses on making full use of saved shot data and uses the adaptive neural-fuzzy system to train the data; the system is based on the Gaussian function. The optimal functional relationship between the shot players, the goalkeeper and shot point can be received. The result shows the efficiency of this method.For the path planning, the paper introduced three common methods, and compared the advantages and disadvantages, and then proposed a path planning method based on S-adaptive genetic algorithm. The main innovation of this method is changed the genetic manipulation, which are crossover probability and mutation probability, so that the probability value can be independently changed according to the hereditary algebras and the individual changes in the group, and saved the much effective individuals. In the same condition, it can make the algorithm convergence faster, and finally get the optimal individual. The path from the S-adaptive algorithm can avoid obstacles in short time, and it is better than traditional genetic algorithms.