Research On The Trajectory Tracking Of Mobile Robots Based On Vision

Trajectory tracking problem has always been one of hot issues in the field of mobile robot research,the robot trajectory tracking algorithm and the position of movement of the signal acquisition are the keys to the effect of tracking technology.In addition,due to the complexity of the actual working environment of mobile robot,obstacle avoiding system can greatly improve the safety of robot in motion control during trajectory tracking.In this thesis,the development process,research status and actual research significance of wheeled mobile robot are introduced.The range of robot vision is expanded under the locating method,which is based on the monocular vision of checker board.At the same time the mobile robot trajectory tracking with obstacle avoidance control strategy is studies.This thesis mainly contributes to the following four aspects.Firstly,This thesis introduces the nonholonomic system and the nonholonomic wheeled mobile robots.Based on the kinematics characteristics of wheeled mobile robot,the world coordinate system and the robot coordinate system is established,and a wheeled mobile robot kinematics model and the model for the trajectory tracking are given.Secondly,for the vision location problem of WMR,based on the introduction of the small hole imaging model and the camera's internal and external parameters,the measurement method of the camera's external parameters based on vanishing point is analyzed and verified by simulation.In view of the problem that image feature point escape from the field of vision,a method to estimate the pose of the robot by multi-group feature points is designed.The problem of signal loss caused by narrow filed of vision is effectively solved.Thirdly,based on the tracking error under the camera coordinate system,trajectory tracking control law is designed by Lyapunov's direct method.Aiming at the problem of simultaneous trajectory tracking and obstacle avoidance,the improved artificial potential field method and fuzzy control method are combined to change the repulsion coefficient and the gravitational coefficient in real time,so as to form virtual targets.Meanwhile,the virtual target point and trajectory tracking control target point are combined together,so that the robot can effectively avoid obstacles and realize the asymptotic stability of trajectory tracking.Finally,the IN-R robot is used to verify the designed algorithm.After parameter adjustment,the robot can track the trajectory and avoid obstacles well,which further confirms the feasibility and good control performance of the algorithm designed in this thesis.