Research And Development Of Strategy And Simulation System Of Robot Soccer

Robot soccer system is a kind of typical multiple mobile robots confrontation system, with the characteristic of high real time request, high dynamic and existing unexpected states. It connects research of high-tech and soccer game, testing many AI theories, methods and technologies via robot soccer games. The robot soccer game has got attentions of researchers in AI and related fields and is becoming a standard problem. The dissertation mainly studies the robot soccer confrontation strategy and some related problems to promote the development of robot soccer, investigates the problem of multi intelligent robots coordination and confrontation, in order to test and develop the relative AI theories and technologies.First, the dissertation formally summarizes the robot soccer's three challenges for the AI subject: real time intelligence, coordination in complex dynamic environments, and antagonizing high intelligent opponents under real time and multi robots environment.Second, it brings out the conception of coarse granularity behavior to divide the strategy problem into four small and simple sub-problems, according to the ideas of IDPI after summarizing many kinds of intelligent robot control structures and analyzing some robot soccer decision-making models. Each sub-problem has solutions from an exsisting research area. Robot soccer strategy architecture based on the coarse granularity behavior and vertical decomposition of intelligent control is proposed and discussed.Third, the dissertation pays attention to two kinds of coordination of multiple soccer robots. Tactics coordination happens in a small local area, with high real time request. Strategetic coordination happens in a large area, with low real time request. To solve the soccer robots coordination problem, the thesis adds an item of operator classification to the Tambe's STEAM model to let the model more fit the robot...