Abstract:
ABSTRACT
ADAPTIVE CONTROL DESIGN FOR CONTROLLING VIBRATIONS
OF 3 DEGREES OF FREEDOM ROBOTIC MANIPULATOR
by
Rostan Rodrigues
Electrical and Computer Engineering
Electronic Engineering Option
California State University, Chico
Summer 2010
In automatic control systems, controlling vibrations of a highly nonlinear and
flexible body is a very challenging task. Many times flexibility in the system causes it to
behave unstable. In this thesis, a PID controller is used to control the vibrations of flexible
tool. Multiple techniques are approached to control the gains of PID controller. A 3
degrees of freedom robotic arm that handles a long, slender, and flexible tool is modeled
into nonlinear dynamic equations by using Newton-Euler method that describes the combined
translational and rotational dynamics of a rigid body.
The simulations of this flexible robotic system are done in Simulink and Matlab.
The simulation results demonstrate the performance differences between various
control schemes. The conventional PID controller have better vibrations settling time and
less steady state error for the first step input as shown in the graphs. But in terms of overall
performance, the gain scheduling PID controller shows better system responses and
steady-state performances for multiple set of control inputs. Because it has the ability to
adjust the gain values according to the changes in system input whereas the conventional
PID controllers has a constant gain value throughout its operation.
Numerical simulation of robot and tool set has been accomplished and results
support the fact that designed gain scheduling PID controllers performs remarkably well
in reducing vibrations and accurate guidance of robot tool tip for tracking various trajectories.
Successful design and implementation of gain scheduling PID controller has
been the main accomplishment of this thesis. Two techniques, self tuning and gain scheduling
are implemented to adapt the controller to system parameter changes. The self tuning
regulator uses system identification in real time and then the PID controller are calculated
on line. Whereas gain scheduling PID controller selects a particular set of gain from
look-up table whenever system parameters are changed by predefined amount. The performance
of self tuning PID controller and gain scheduling PID controller are investigated.
