Research On Obstacle Avoidance Of Wheeled Robot Based On Visual Positioning

In this thesis,the problem of the stabilization of a wheeled mobile robot(WMR)system based on vehicle vision positioning in a complex environment is studied.On the basis of the traditional artificial potential field method,an algorithm to realize the obstacle avoidance of mobile robot is proposed in the process of system stabilization.A repulsive force factor is introduced when the robot detects obstacle during the movement,and the obstacle avoidance is realized by the acceleration provided by the repulsive force factor and the current robot velocity.At the same time,the visual positioning control system and the encoder positioning control system are switched to realize the global obstacle avoidance control of WMRThe main contents of this paper include:First,based on the kinematics characteristics,a kinematics model of wheeled mobile robot is given.On the IN-R WMR experimental platform,a real-time pose acquisition algorithm of robot is realized by using 2D triple focus tensor method.Secondly,The traditional artificial potential field method is improved.Based on the stabilization control of wheeled mobile robot,a repulsive force factor is introduced when the robot encounters obstacles,and the obstacle avoidance is achieved by the interaction between the acceleration provided by the repulsive force factor and the current robot speed.Then,a controller is designed to asymptotically stabilize the mobile robot with the Lyapunov function method.Furthermore,in order to solve the problem of unknown disturbance and parameter uncertainties of robot in actual motion,a sliding mode controller is designed for the encoder positioning control system and an adaptive sliding mode controller is designed for the visual positioning control system,which enhances the robustness of the system.Finally,by using checkerboard detection based on OpenCv library function,the automatic search of the target by the wheeled mobile robot is realized,and then the target switching of the robot is realized when the tracked corner is offset.On this basis,through the design of angle offset identification area and the virtual view angle constraints in the camera image plane,the switching between visual positioning motion control system and encoder positioning motion control system is realized,so as to solve the problem of the loss of visual orientation when using visual positioning.Experimental results on IN-R experimental platform prove the effectiveness of the above method.