Dynamic modeling, experimental identification, and active vibration control design of a 'smart parallel manipulator'

This thesis addresses dynamic modeling of a high-speed flexible-link planar parallel manipulator, experimental identification and active vibration control of vibration in flexible linkages on the parallel manipulator using distributed smart material components. Planar parallel manipulators undergo significant linkage vibration caused by large inertia forces associated with high accelerations and high-speed motion. Settling time of unwanted vibration slows down the operational speed of parallel manipulators and residual vibration adversely affects the trajectory control accuracy. A "smart parallel manipulator" is proposed to actively control vibration in intermediate linkages using distributed arrays of surface-bonded Lead Zirconate Titanate (PZT) patch sensors and actuators so that operation speed of the manipulator is greatly increased without sacrificing trajectory control accuracy. Modeling and control of the smart parallel manipulator is characterized by (i) closed-loop architecture of the parallel manipulator with multiple kinematic chains, (ii) coupling of nonlinear rigid body motion and flexible deformation, and (iii) configuration-dependent vibration characteristic of the parallel manipulator.; In this thesis, a finite element (FE) model of flexible linkages is developed, where the coupling of translational motion, rotation, and flexible deformation and coupling of PZT sensors and actuators with the host structures are incorporated. A substructuring modeling procedure is presented to develop dynamic models for closed-loop flexible mechanisms that are capable of modeling essential dynamic behavior of the parallel manipulator with moderate model order. Motion control and active vibration control simulation numerically demonstrate the concept of smart structure control in the parallel manipulator. An experimental identification procedure using various combinations of transducers has been developed to identify linkage vibration characteristics when the planar parallel robot is in both a stationary state and during rigid body motion. Configuration dependency of linkage vibration characteristics is investigated through intensive identification experiments. An active vibration control algorithm in the modal space is developed and implemented to control configuration-dependent linkage vibration. Vibration caused by linkage flexibility is substantially reduced.