Research And Implement Of RoboCup3D Simulation Robot

RoboCup (Robot World Cup) is Robot World Cup Soccer,which take MAS(Multi-Agent System) and DAI (Distributed Artificial Intelligence) as researchbackground.The main purpose is to promote the research and development of MASand DAI by providing a standard and evaluable platform.RoboCup3D is performed in a standard environment provided by RoboCupcompetition.We have created a team call as YuShan3D and desinged a worldmodel,individual skills, information processing and positioning strategies to make theagent be able to compete in the game.Research results can be applied in manyfields,such as intelligence,space exploration,search and rescue and industrial roboticsso on.This paper designs and implements YuShan3D soccer team,The specific work ofthis paper as follows:1. Accurate world model:World model was described the internal state ofsimulation soccer game and divided into four layers:information storage layer(Basic),updating layer (Update),predicted layer (Pridict) and high-level layer(HighLevel).2. Precise localization:Agent’s self-positioning and locating other objects on thefield was with the help of the vision sensor located in the head. In order to get rid ofthe noise in the game，using an extended kalman filter trace of the soccer ball.3. Improved individual skills:the development of agent’s motions wereimplemented by the GUI debugger.The walking behavior was generated throughonline way.4. Efficient high-level strategy: A initial formation was given in YuShan3D,eachplayer will be assigned a specific role according to the role assigned function and willswitch dynamically when game environment has changed.We applied these research results to our YuShan3D team.In the game,our teamhas shown excellent score ability,and we have won the first prize in the RoboCupChina Open2013and second prize in the RoboCup China Open2012.