Research On Multi-sensor Fusion Localization Method Of Indoor Mobile Robot Based On Two-dimension Code

In recent years,with the in-depth development of robot and self-driving technology,the research on robot localization technology has gradually become a hot spot.At present,most indoor mobile robots estimate their own location through all kinds of sensors installed on the robot,but affected by the accuracy of the sensor,if only rely on the information of a single sensor,it will affect the localization accuracy to a certain extent.To improve the localization accuracy,an algorithm based on iterative extended Kalman filter(IEKF)fusion of two-dimensional code and odometer is proposed in this paper.In the case that the odometer cannot be used,the location of the robot is estimated by using QR code,lidar and IMU sensors and information fusion with the help of extended Kalman filter(EKF)and adaptive Monte Carlo localization algorithm(AMCL).The specific research contents are as follows:The AprilTag is used as the artificial road sign to carry out the research on mobile robot aided localization.Considering that some of the Tag images are dark,for improving the recognition accuracy,the relevant image processing techniques are used to process the images captured by the camera to obtain an ideal Tagimage for subsequent localization work.In order to enable the mobile robot to follow the predetermined route even when the two-dimensional code cannot be recognized,IEKF is taken as the main algorithm to design the algorithm for the odometer data fusion two-dimensional code information.Considering that some mobile robots cannot install odometer sensors,a new sensor data processing scheme is proposed,that is,the motion data of the robot is extracted from the lidar by using the RF2O(Range Flow-based 2D odometry)algorithm,and the inertial measurement unit data and the calculation results of the RF2 O algorithm are fused with the EKF algorithm to locate.In view of the problem that the poor feature extraction effect of lidar in the case of a single environmental feature will lead to a poor positioning result of RF2 O algorithm,this paper proposes to use AprilTag's information to modify the fusion location results of EKF in the adaptive Monte Carlo localization algorithm in order to get a more accurate location.Finally,the mobile robot platform and Gazebo simulation environment are designed and constructed,and on this basis,the actual and simulation localization experiments of the robot are carried out.The results show that the localization method proposed in this paper has good practicability and high accuracy.