A Localization System With Global Visual Fusion Based On Double AR Code

This topic takes the global visual positioning as the core and integrates multiple sensor information,and uses the robot operating system(ROS)as the software framework.This paper takes the accuracy and stability of the positioning of mobile robots in SLAM research as the starting point,in order to reduce the influence of the inevitable factors in the environment on the positioning of mobile robots,so that mobile robots can work stably and safely in the workshop and work room.Break through the bottleneck of mobile robot industrial application of mobile operation.The main contributions of this paper are as follows:Firstly,comparing the characteristics of the positioning scheme of different camera installation positions in the indoor environment,the mobile robot positioning system based on global vision is designed by using monocular vision and two-dimensional code.By attaching and fixing the camera to the top of the indoor environment,the QR code is pasted on the robot,so that the camera can stably recognize and track the mobile robot,and track the robot pose and generate the motion track in real time while the robot is moving forward,and use the robot.The relative error of the upper two-dimensional code is analyzed to analyze the error of the global visual positioning system.Secondly,based on the analysis of the global visual positioning system error,a positioning fusion method combining global visual positioning with the odometer pose estimation results carried by the robot is proposed.Aiming at the measurement discreteness of the global vision sensor in the traditional measurement system fusion scheme and the local failure of the indoor observation result caused by the illumination effect,and the insufficient optimization of the fixed variance of the traditional fusion scheme,the calculation of the variance of the two-dimensional code is proposed.The confidence of the global visual positioning system.In addition,through the in-depth analysis of the odometer information-based ranging error source and its advantages,the odometer's positioning information is added to the Kalman algorithm(EKF)for local fusion filtering with the global visual positioning system.The multi-sensor measurement makes the robot positioning model more accurate,so as to reduce the positioning error of the robot and improve the accuracyand stability of the robot positioning.Thirdly,aiming at the basic application requirements of mobile robots in the positioning and navigation of indoor scenes,this paper proposes a mobile robot navigation strategy based on global visual positioning.The strategy combines the absolute positioning acquired by the global visual positioning system with the relative positioning of the odometer information to obtain the obtained robot pose,and transmits it to the mobile robot chassis for stable and reliable navigation.Realize the accurate and stable positioning of autonomous mobile robots to achieve point-to-point transportation and specify line cruise.Finally,in this paper,experimental scenarios are designed in an indoor environment,and multiple experiments are performed using a robot.The experimental results show that the proposed global vision fusion Localization system based on double AR code is effective and reliable.