System Research And Design Of Identifier Recognition And Path Tracking Of Vision-guided Mobile Robot

Promoted by Industrial 4.0 and Made in China 2025 policy,domestic and overseas universities and enterprises gradually focus their research on robots.With the improvement of software and hardware performance,such as computers and sensors,functions of mobile robots become more comprehensive and full-automated.As the "eyes" of robot,visual navigation is always a hotspot in the study of robot guidance,the key technology of which is path recognition.Improving the performance of path recognition and path tracking control of mobile robots is beneficial to realize higher precision motion,track more complex routes and accomplish more difficult tasks.This paper took the vision-guided mobile robot as the research object.The vision-guided system for robot was designed and built.The correction of distortion images was achieved by modeling and calibrating for camera.The preprocessing algorithms of navigation belt image and identifier image was studied,including grayscale,filter,segmentation,morphological enhancement,edge detection and region selection.Through contour edge scanning method proposed by this study,edge lines of navigation belt image was extracted,and its center line was fitted based on least square method.The improved Hu moment algorithm was utilized to extract the identifier features,and the machine learning algorithm was used to recognize identifier.The equation of the path tracking control system was obtained by establishing the kinematics model of the vision-guided mobile robot and analyzing the mathematical relationship between the positional deviation and angular deviation of the two driving wheels and the control voltage.A fuzzy controller was designed for the kinematics model based on the fuzzy logic theory,and the performance of it was compared with the classic PID controller.The vision-guided mobile robot was tested to verify the accuracy and real-time performance of the visual image processing algorithm and the corrective performance of the path tracking controller.The test results indicate the followings.Firstly,compared with the traditional algorithm,the proposed contour edge scanning method combined with the least squares method to fit the centerline took 22 ms.The processing time of improved Hu moment algorithm is 2.49 ms,which is shorter than 4.81 ms of the original Hu algorithm,and the precision of improved Hu moment algorithm is higher.The performance of the two improved algorithms satisfies the real-time requirements and accuracy requirements of the vision-guided mobile robot system.Secondly,path tracking controller,which is based on fuzzy logic theory,can accurately and rapidly track the straight routes,arc routes and mixture routes,and controls angle deviation to ±4° and position deviation to ±6 mm.It shows that the path tracking controller has high performance as well as great robustness.This controller can be applied to the mobile robot system of different speed and is of great significance for engineering application.