[Th015]
Citation:
D.Y. Hwang, 'Teleoperation Performance with a Kinematically Redundant Slave Robot,'
Ph.D. Dissertation, University of Washington, Department of Electrical Engineering, December 1995.
Abstract
In this dissertation, two topics are presented: 1) scaled
teleoperation system stability, 2) teleoperation with kinematic
redundancy. When the scaled teleoperation is used, the stability
issue becomes important due to the high gains in position or force
scaling. We used two experimental methods to identify the stable
region of position and force scaling factors of a one-axis system
haveing 4000:1 force scaling capability. We used a simple classical
teleoperation control law. An indirect closed loop approach with ARX
technique was used to model the nonlinear slave system with a flexible
printed circuit cable. The stability regions were identified by
experiments and simulations. Then the slave system was physically
linearized by replacing the flexibile printed circuit cable with thin
wires. The linear case stability regions were identified. In
addition, the contact stability region and sensitivity analysis were
studied.
Kinetic energy minimization was used to resolve the
redundant joint space solution in the early literature work by Whitney
in 1969. Kinetic energy optimization has been used for local and
global redundancy control. However, the further effects of kinetic
energy on teleoperation performance have not been studied. This
section shows the effects of kinetic energy level on teleoperation
performance. First, we expressed an inertia weighted pseudo-inverse
solution corresponding to a general level of kinetic energy by using
the gradient projection method. Second, we added joint limit avoiding
by using the null space of the main task. Then we derived three
kinematic redundancy control modes having different kinetic energy
levels using the property of the generalized inverse matrix: "Whitney
Mode", "Intermediate Mode", and "Least Squares" mode.
To validate
our algorithms, we performed autonomous robot tests and teleoperation.
The tradeoff between kinetic energy level and joint limit index was
clearly shown in the autonomous test. For teleoperation, 4 tasks and 7
indices were defined. A 3-DOF pen-based master and 5-DOF Mini Direct
Drive Robot were used. Tasks were x-y-z positioning and contact,
giving 2 DOF kinematic redundancy in the slave robot. Overall, the
"Whitney Mode" showed the best performance, while the least-squares
mode showed the worst performance. In addition, the subjective quality
ratings of inertia force, bias force, noise force were made by the test
operators. "Whitney Mode" showed the lowest perceived level of noise
force.
["I would like a hard copy of this report"]
[Copyright]
[HELP!]
Updated: Tue Jul 15 23:54:51 2008