ADAPTIVE CONTROL DESIGN FOR CONTROLLING VIBRATIONS OF 3 DEGREES OF FREEDOM ROBOTIC MANIPULATOR

ABSTRACT \nADAPTIVE CONTROL DESIGN FOR CONTROLLING VIBRATIONS \nOF 3 DEGREES OF FREEDOM ROBOTIC MANIPULATOR \nby \nRostan Rodrigues \nElectrical and Computer Engineering \nElectronic Engineering Option \nCalifornia State University, Chico \nSummer 2010 \nIn automatic control systems, controlling vibrations of a highly nonlinear and \nflexible body is a very challenging task. Many times flexibility in the system causes it to \nbehave unstable. In this thesis, a PID controller is used to control the vibrations of flexible \ntool. Multiple techniques are approached to control the gains of PID controller. A 3 \ndegrees of freedom robotic arm that handles a long, slender, and flexible tool is modeled \ninto nonlinear dynamic equations by using Newton-Euler method that describes the combined \ntranslational and rotational dynamics of a rigid body. \nThe simulations of this flexible robotic system are done in Simulink and Matlab. \nThe simulation results demonstrate the performance differences between various \ncontrol schemes. The conventional PID controller have better vibrations settling time and \nless steady state error for the first step input as shown in the graphs. But in terms of overall \nperformance, the gain scheduling PID controller shows better system responses and \nsteady-state performances for multiple set of control inputs. Because it has the ability to \nadjust the gain values according to the changes in system input whereas the conventional \nPID controllers has a constant gain value throughout its operation. \nNumerical simulation of robot and tool set has been accomplished and results \nsupport the fact that designed gain scheduling PID controllers performs remarkably well \nin reducing vibrations and accurate guidance of robot tool tip for tracking various trajectories. \nSuccessful design and implementation of gain scheduling PID controller has \nbeen the main accomplishment of this thesis. Two techniques, self tuning and gain scheduling \nare implemented to adapt the controller to system parameter changes. The self tuning \nregulator uses system identification in real time and then the PID controller are calculated \non line. Whereas gain scheduling PID controller selects a particular set of gain from \nlook-up table whenever system parameters are changed by predefined amount. The performance \nof self tuning PID controller and gain scheduling PID controller are investigated.