Research And Implement Of Method Of Localization For Four-legged Robot

Robot soccer is a hot spot in current intelligence control field. It provides a goodplatform for autonomous mobile robot's research. In the context of the RoboCupFour-Legged League, this paper researches the technology of robot localization,furthermore, implemented the localization module of the match system, under theexperiment platform of AIBO ERS-7, which is produced by Sony Corporation.In the first part, it researches the theory of Bayesian Localization method, andgives the algorithm, besides, it presentsmainly on Monte Carlo Localization method.In the second part, in the context of the RoboCup Four-Legged League, underthe Sony Four Legged Robot AIBO ERS-7 platform that used in the math, thelocation system is implemented, on the base of Monte Carlo method that wementioned above. And the structure of the location system is also given. Samplingfrom the uniformity distribution is implemented. And it can use the movementinformation to update the sample set. According to the different environmentinformation that the robot observed, it makes analysis and gives the algorithm. Inaddition, it gives the linear time re-sampling method to resample the sample set. Inpoint of the ball location tracking problem in the match, the system make a specialdisposal of the moving ball localization, it introduce the ball speed information intothe re-sampling method, to enhance the veracity of the ball localization. Finally,through the experiment of the implemented location system, the validity and veracityof the system are conformed.In the finality, the localization achieved by the experiment proves the validityand accuracy of the system.