| This thesis mainly studies the trajectory tracking problem of controller design fornonholonomic mobile robots. At first the basic model of the nonholonomic mobilerobots and the related basic thory were discussed, such as the camera imaging modeland image processing theory and so on, which are prepared for the trajectory trackingcontroller design of nonholonomic mobile robots with stereo vision servo. At first,based on these, the model which is based the position vision servo (PBVS) wasestablished, and controller of the target trajectory tracking was proposed. Meanwhile,the controller’stability was proved and simulated in the computer. To improve thecontroller, a position error tracking controller was proposed. Secondly, the model whichis based the imageEstablishing a model which is based the image vision servo (IBVS)was established, and controller of static target tracking was proposed. The strict mathprove is used to ensure the stability the controller, and the other image based positioncontroller is designed to ensure the target is always in the common scope of twocameras.Thirdly, as for the case that the target is near to the robot, the goal will probably not located in the public view, and it will be difficult to apply the stereo vision servomodel, a monocular visual servoing feedback controller was designed which can make the robot track the target in the situation that the depth is unknown but it has upper bound. Finally, since true performance of robot tracking cannot be completely shown by the theoretical model analysis and computer simulation, a large number of experiment had been done, it mainly included: stereo vision calibration experiment, recognition of targetin the image and stereo vision camera reappear experiment in3D, robot self–localization experiment, target’s relative position recognition experiment, experiment of map automatic generation by robot real-time operation, relevant controller performance verification and robot visual integrated navigation experiment. |
PDF Full Text Request