Research On Position Control Of Mobile Robot Based On Uncalibrated Visual Servoing

With the development of computer vision and robot technology,visual sensors have been widely used in the field of mobile robots.Mobile robot visual servoing refers to a method of directly controlling mobile robot based on visual feedback.Traditional visual servoing generally relies on the calibration of camera parameters,which is difficult to handle a complex and changing environment.Therefore,based on feature points,an uncalibrated visual servoing system is designed to achieve high-precision position control of mobile robots.Firstly,the overall design scheme is determined.The mobile robot platform and wireless communication module are designed.The accuracy of the open-loop motion control is verified when the robot receives remote commands.Secondly,robust feature point pattern with recognition and tracking algorithms are proposed.Aiming at the requirements of visual servoing input,the mobile robot feature points including positioning points and coding points are designed.In static image,the contour-based feature-point-recognition method is proposed.In dynamic images,the Median-Flow tracking algorithm is used to track the mobile robot feature point.The high-precision recognition and real-time positioning algorithms are verified through experiments.Then,the uncalibrated visual servoing algorithm is studied to realize the stabilization control and trajectory tracking control of the mobile robot.In order to improve the accuracy when the camera is arbitrarily placed,a static depth estimation method is proposed.The control accuracy with proposed algorithm is verified through experiments.Finally,a trajectory planning method is designed based on the success of trajectory tracking control.Object recognition algorithm is designed based on target detection and image segmentation.The map is generated by grid map method.Then A-star algorithm is used to achieve obstacle avoidance trajectory planning of mobile robot.The image-stitching algorithm is adopted to realize the migration movement of the mobile robot between two images with overlapping areas.The obstacle avoidance and migration trajectory-planning algorithm are verified through experiments.