Nonlinear Control Of Image-based Robotic Visual Servoing

As the development of science and technology, we have further studies on robot,visual robot gradually apart from other types of robots with its large amount ofinformation, high accuracy, etc., and become a hot topic in robotic research. But most ofresearch in visual servo, directly or indirectly, assume that the robot joint speed is known,in practice, to get the robot joint speed often need the speed sensor, on one hand, this willincrease the actual cost of commercial applications, on the other hand, the use of thesensor may introduce noise and affect the control accuracy. To solve this problem, we willstudy to design the I&I observer to estimate the speed of the robot joints, and designcontrol scheme for the visual servoing system with considering the unknown cameraparameters. In addition, we will use the limited-time control theory to regulate the visualservo robot, for its advantage of noise immunity, this will enhance the robustness of thesystem. The concrete work is as follows:First of all, we have learnt the research background and research status in the field ofvisual servoing robot, as well as the basic theoretical knowledge for the research such asrobot dynamics, robot kinematics, transformation between the coordinate system and thethe basic theory of I&I theory. On this basis, we will propose the problem of visualservoing with the joint speed unknown.Secondly, as the assumption referred before, we will design the I&I observer basedon the I&I theory. Compare with the existing observer, this observer can achieve systemglobal asymptotically stable. A PD-like image-based visual servo control algorithm will begiven based on the speed observer, and the stability analysis of closed loop system andsimulation are also given.Again, as the work done above haven’t consider the condition that the cameraparameter are unknown or the camera are uncalibrated, in this part, we will not onlyconsider the robot joints’ velocities are unknown, but also suppose that the camera areuncalibrated. We will use the depth-independent matrix to linearization the unknowncamera parameters from Jacobian image matrix, on this basis, we will design an adaptivecontrol algorithm to achieve the objective of visual servoing. At the last, the Lyapunov stability analysis and simulation results will be given.Finally, as we known, finite-time control is a robust nonlinear control method, butmost of the research results of visual servoing did not consider the finite time controlmethod as visual servoing robot system is a complex nonlinear system, We will design avisual servoing finite-time controller. The simulation results proved that this scheme canenhance the robustness of the system.