Research On Indoor Positioning And Motion Control Of Omnidirectional Mobile Robot Based On ROS

Owing to the excellent mobility and passability,omnidirectional mobile robot has broad application prospects in industrial logistics,production assembly,etc.This dissertation focuses on the precise indoor positioning and motion control of omnidirectional mobile robot.Several key technologies are researched and implemented,including robot kinematics modeling,inertial measurement unit dynamic error compensation,magnetic compass yaw correction,and multi-sensor information fusion based on extended Kalman filter.The major contributions of this study are as follows.The dissertation establishes the kinematic model of omnidirectional mobile robots,including the kinematics equation and control model of the four Swedish wheels mobile robot system,and the sensor observation models of the motor encoder,measurement wheel encoder and inertial measurement unit are also proposed.To reduce the adverse effects of the static error and zero drift of the gyroscope in the IMU on the attitude estimation,the data fusion of the gyroscope,the magnetic compass and the accelerometer is carried out using the Mahony complementary filter algorithm.Furthermore,a dynamic gyroscope error compensation algorithm based on the non-rotating state discrimination is proposed.The compensation algorithm can effectively compensate the zero drift of the gyroscope.To address the issue of yaw angle error of magnetic compass under ambient magnetic field disturbance,a yaw angle correction method and a magnetic field interference discriminating method based on rotating magnetic compass are introduced.Based on elliptical fitting,the hard iron and soft iron corrections as well as the method that estimate the magnetic field distortion are combined to achieve comprehensive correction of yaw.Based on the dynamic error compensation of gyroscope and the yaw angle correction of magnetic compass,the positioning data from the driving wheel encoder and the combination of measurement wheel with inertial measurement unit are fused to improve the robot's relative localization accuracy using the extended Kalman filter algorithm.A four Swedish wheeled omnidirectional mobile robot prototype is developed,and a Uniform Robot Description Format file for the mobile robot is designed for the ROS integration.The onboard control system as well as the ROS application node program are implementated to achieve system function.The tracking control model and the deviation model of the omnidirectional mobile robot on the basic trajectory are studied,and the tracking control of the straight line,circular arc and cubic curve is realized.Related experiments show that the trajectory tracking accuracy can meet the general application requirements.