Research On Self-location And Navigation Of Mobile Robot Based On Artificial Landmark

The self-location and navigation function of Mobile Robot is always one of research hotspot for scientific research workers.Landmarks which have certain contrast ratio with the environment are the foundation of Mobile Robot self-locating and navigating.They naturally become the first selection for Mobile Robot to realize the self-location and navigation because of their color and shape easy to set,flexible to layout,easy to extract and so on.In this thesis,being aimed the complicated environments such as the roof structure in the Lab,a kind of vision navigation system for Mobile Robot based on artificial landmarks was designed.In order to locate and navigate Mobile Robot,the following research work has been done.Firstly,the kinematics analysis of wheeled Mobile Robot differential driven was taken and the vision navigation system for Mobile Robot based on artificial landmarks was designed.According to the hole imaging principle of camera and the characteristics of vision navigation system for Mobile Robot,the model of Mobile Robot localization based on monocular vision was established.Secondly,taking into account the obvious contrast between the artificial landmark and the background,feature extraction was completed by using HSV in order to minimizing the interference of highlights such as daylight lamps as much as possible.The artificial landmarks' central points has been obtained by fitting contours according to regional area,contours' perimeter and circularity in order to eliminate the non-feature contours.Thirdly,in view of the characteristic of camera's imaging,the calibration of the camera has been completed based on OpenCV software.The camera's internal matrix and distortion vector were obtained.Accordingly,the camera parameters such as minimum degree of lens and viewing angle were obtained through internal matrix and the basic parameters of the camera.The vision navigation system for Mobile Robot based on artificial landmarks was established according to the calibration results of the camera and the other parameters known.Fourthly,the positioning algorithm for Mobile Robot was deduced according to the vision navigation system designed above.In order to improve the positioning precision of Mobile Robot,the calibrating principle for (?) position and orientation(named as the pose)of camera relative to Mobile Robot was analyzed and it has been calibrated.At the same time,the actual location of all artificial landmarks has been calibrated.Fifthly,according to the hardware foundation of XAUT·AGV100 used in experiment,self-location and navigation control system for Mobile Robot was designed based on LabVIEW software.And the trajectory tracking experiment that Mobile Robot moves along trajectory like "L" within visual navigation system platforms has been done.The main reasons for the location error for Mobile Robot were analyzed according to the actual position and orientation of Mobile Robot at the target point.It is verified by the result of experiment that the vision navigation system for Mobile Robot based on artificial landmarks designed can be used Wheeled Mobile Robot to self-locating and navigating.And the algorithm for extracting the feature of the artificial landmark is correct.The algorithm for location Mobile Robot is correct.