Research on Localization Technology for Mobile Robots

Abstract

By the testing platform base on two-wheel differential drive of three-wheel robot,the robot positioning system was composed of encoder and gyroscope,then the robot kinematic equation was established.The two sensor's data were fused using Kalman filter,and the data was regressed by the weighted least squares support vector regression for construction of prior regression decision function.By using the prior regression decision function to improve accuracy of the angular rate of gyroscope,the information of location and pose of the robot were acquired.