Robot Manipulator Five-degrees Of Freedom Uncalibrated Visual Servoing Based On Human-simulated Intelligent Control

Robot manipulator visual servoing combine s robot manipulator control and visual perception, and robot manipulator can locate and track the target in the workspace under the guidance of the visual feedback. Robot application based on the visual perception is no longer limited to repeat the work in a fixed and known work environment, and it expands the using range of the robot manipulator.In the camera model and the robot model uncalibrated case, robot manipulator uncalibrated visual servoing designs the controller according to image feature error in the image plane, and the image fea ture change will be mapped to robot manipulator motion in the Cartesian space by using the visual mapping model, then direct drive robot manipulator moving to the desired posture until the image feature error converges to zero. Due to uncalibrated visual servoing doesn’t establish the precision mapping model between the image features change and robot manipulator motion, and the robot manipulator is a highly nonlinear and comple x time-varying dynamical system. Therefore, the essence of the robot manipulator uncalibrated visual servoing is a typical unmodeled dynamics nonlinear control problem, and it is also the research focus and emphasis of the robot intelligent control field.Human simulated intelligent control(HSIC) theory design the controller by imitating the control behavior and experience of human, which is effective for the controlled object model is unknown and complex nonlinear control problems. This paper proposed the robot manipulator five-degrees of freedom uncalibrated visual servoing method based on human simulated intelligent control theory, and the main contents of this paper are organized as follows:①the robot manipulator five-degrees of freedom uncalibrated visual servoing method based on humanoid intelligent control theory was proposed. The binocular vision Jacobian matrix based on the point and line feature is approximate to descript the dynamic hand-eye relationship between the image feature changes and robot posture changes, and then divide image eigenmodes by Combining with the prior knowledge of the robot and camera system. Moreover, the human simulated intelligent multimodal visual controller was designed to solve robot manipulator five-degrees of freedom uncalibrated visual servoing on the basis of the kalman filter estimation updating image Jacobian online, as well as the qualitative decision and quantitative control two-order mapping model of the human simulated intelligent control. Finally, this paper obtain the stability condition of the human simulated intelligent multi- modal visual servoing controller.②The robot manipulator uncalibrated visual servo simulation platform was established by combining with the Puma560 robot manipulator model, the industrial camera model and the Simulink model. Moreover, the robot manipulator five-degrees of freedom uncalibrated visual servoing simulation experiment was conducted.③The robot uncalibrated visual servoing physical platform was established based on the robot manipulator and two industrial cameras. In addition, the robot manipulator five-degrees of freedom uncalibrated visual servoing comparative experiment was conducted between the proposed method and the method of two point image feature based on the visual mapping model of the Jacobi matrix.Simulation and experiments verify the effectiveness and practicality of the proposed method in the robot manipulator five-degrees of freedom uncalibrated vision positioning. The physical contrast experimental results show that the convergence speed of the proposed controller in this paper is quicker and robot manipulator trajectory is smoother. In addition, it obtained better control effect.