Position-based Visual Control And Target Tracking

Visual servoing systems based on position transmits the perceived environmental information into motion control systems,and under the condition of no contact accomplishes target localization and tracking.The servoing system mainly includes two parts,namely,visual systems and control systems.Both are connected with the relative position estimation algorithm.Convert image information into the location of the world coordinate system.By adding the visual sensor to the mechanical arm,the environmental awareness is enhanced and the improving the value of the mechanical arm.The thesis first expounds the visual servoing control system of the research background,significance,and the development direction of the future,summarizes the present research achievements of the sliding mode control algorithm,and contrasts the PBVS and IBVS.Beyond that,it detailedly introduces the knowledge of kinematics and dynamics.It uses the robot-toolbox to simulate kinematics and dynamics of the puma560 and calibrates the camera offline at the same time.It comes up with the relative pose estimation problem based on PBVS.As for plane moving objects based on kalman filter algorithm it estimates the Jacobian matrix online,and eventually completes the target tracking and verifies it by simulation experiments.Sencond,the sliding mode variable structure control algorithm to realize target tracking is proposed.In order to speed up the state points to reach the equilibrium,we designs the sliding mode reaching law and the sliding mode surface as follows: 1.Choose composite reaching law in combination with uniform functions,exponential functions and power functions,to guarantee the quickness of the state points reaching the sliding surface.Use the piecewise function rather than the symbolic function,to weaken the chattering problem.2.Add the nonlinear term to the linear sliding mode surface,and ensure the quickness of the sliding mode motion.Finally take the outside interference of servo system into consideration,design a nonlinear observer based on the principle of linear observers.The organic combination of sliding mode controlling and the nonlinear observer can get the unknown disturbance estimated and achieve the goal of accurate tracking at the same time.