| Omni-directional mobile robots with high efficiency,reliability and flexibility tent to be a hotspot in research field.Accurate navigation in mobile robots would implement extensively,and odometry is a critical method of relevant location,so improving precision of odometry to build up reliable autonomous system would be meaningfulWith reference to a goal of improving precision of odometry,the main body of research as following,(1)To obtain flexibility of omni-directional mobile robots,this research adapt decoupled powered caster wheels as the main drive.Using different parameters to build up a kinematic model of vehicle velocity at the same coordinate,and the matrix condition number is introduced to analyze the stability of the kinematics matrix.In order to ensure the continuity of acceleration in both workspace and joint space at the same time,the fifth-order polynomial trajectory planning for robot is employed.Based on the concept above,a master and slave motion control system is established by open hardware framework.The master computer is an industrial one running ROS,and the Galil motion controller that performs motor speed interpolation as slave system.(2)In order to improve the precision of odometry caused by system error,an odometry parameters calibration method for the Powered Caster Wheeled Mobile Robot(PCWMR)is proposed.By analyzing the kinematics matrix,systematic error source of the PCWMR odometry is equivalent to the caster radius,the caster offset and the installation distance of the caster shaft to the center of the platform.In order to reduce the coupling between all of the odometry parameters,the calibration method is designed into two trajectories based on the angle coordination of caster: the first trajectory drives the casters to roll,and the second trajectory drives the casters to steer and roll simultaneously.Then environment model is established,and a simulation is completed by the least squares algorithm to make the odometry close to real value.Finally,the Qualisys optical motion capture system and joint information acquisition system is employed to verify that if the calibrated parameters could improve the odometry precision.(3)In order to reduce the influence of non-systematic error on the accuracy of odometry,a joint space multi-sensor fusion method with IMU and encoders based on Kalman filter is proposed.The MATLAB is used to simulate the IMU and encoders data according to the actual situation,the simulation results show that the sum of squares of the steering joint angle error,the motion control error and the odometry error are reduced more than 95%,97% and 96% respectively. |
PDF Full Text Request