Research On Visual Road Detection And Trajectory Tracking Control Of Security Robot

Security work is indispensable,in the plant,park,community and other scenes.Using security robots instead of manpower for security inspection can effectively deal with the shortage of human resources in the environment of aging population.As the key technology of security robot autonomous cruise,visual navigation has the advantages of low cost and rich information.However,at present,there are some problems,such as poor real-time performance,low accuracy of road detection and poor trajectory tracking accuracy,which can not meet the requirements of robot autonomous cruise.Therefore,it is necessary to carry out the research on visual road detection and trajectory tracking control of security robot,so as to improve the ability of road detection and trajectory tracking in the process of security robot performing tasks,and provide a theoretical basis for the application of security robot.Firstly,starting from the functional requirements of security robot,the robot system is divided into five parts:drive control subsystem,emergency disposal subsystem,communication subsystem,perception subsystem and main control subsystem.This thesis analyzes and designs each subsystem,completes the development of security robot prototype,and develops a microwave-5G/4G autonomous switching wireless communication system for its communication subsystem.The prototype provides a verification platform for the research of visual road detection and trajectory tracking control.Secondly,for the problems of poor real-time performance and inaccurate road boundary recognition of security robot visual road detection technology,the road detection algorithms is improved.Through binocular camera calibration,the image distortion is reduced and the input image quality is improved.In the stereo matching stage with large amount of calculation and time-consuming,GPU acceleration scheme is adopted to improve the speed of disparity map calculation.In order to improve the accuracy of road detection,a multi clue fusion road detection algorithm is proposed.Based on the road detection of V-disparity map,the gray value is adaptively segmented,and finally the accurate road detection area is obtained.In addition,the identified road boundary is sampled at intervals,and the quintic polynomial is used for road fitting as the expected trajectory of the robot.Through the dataset test,the accuracy of the road detection algorithm in this thesis is verified by comparing with the unimproved algorithm.Subsequently,in order to make the robot travel stably along the expected trajectory,the research on robot trajectory tracking control is carried out.The kinematics and dynamics models of the robot are constructed,and the accurate transfer function expression of the robot model is obtained by the method of system identification,which is verified by experiments.The longitudinal control is relatively simple,so the double PID controller is used for speed regulation.For lateral control,the LQR control algorithm of adaptive preview distance is proposed.The fuzzy control method is used to generate the preview distance according to the current vehicle speed and road curvature.The trajectory tracking simulation is carried out in Matlab/Simulink to verify the effectiveness of the trajectory tracking algorithm proposed in this thesis.Finally,by building the security robot experimental platform and software environment,the basic performance experiments are carried out to test its motion performance,control performance and communication performance.The experiments of road detection and trajectory tracking are carried out in outdoor real scene on the security robot prototype.The experimental results show that the road detection and trajectory tracking algorithm proposed in this thesis can achieve the effect of accurate recognition and tracking,and the security robot system can achieve the expected indicators and meet the application requirements of security robot.This thesis is supported by Shandong major science and technology innovation project(No.2019JZZY010112),key research and development program(2020JMRH0202),major industrial research projects in Shandong Province for the conversion of old and new kinetic energy(2021-13).