M.S. Thesis Presentation by Vincent Chan
Friday, March 23, 2001

(Dr. Imme Ebert-Uphoff, advisor)

"Singularity Analysis and Redundant Actuation of Parallel Manipulators"


Parallel manipulators are a classification of robotic manipulators that have closed-loop kinematic chains.  They are commonly made up of a mobile platform that is connected to the base by a number of legs.  Parallel manipulators are gaining acceptance in industry because they are well suited to applications that require moving a large or heavy payload, and they also excel in the areas of stiffness, speed, and accuracy.  Singular configurations of parallel manipulators exist which can result in dangerous situations where the manipulator is unable to sustain a force or wrench applied at the robot end-effector.

Singular Value Decomposition (SVD) is employed to analyze the singular configurations of parallel manipulators.  The SVD is a universal tool that can be applied to any manipulator with knowledge only of the Jacobian matrix.  In general, the SVD is not scale-invariant.  However, in singular configurations, it is appropriate to use the SVD to calculate the nullspace of the Jacobian, since the results in this case are truly independent of scale.  Analysis of the singular configurations using the SVD can provide a better understanding of the problem and aid in the design of a redundant actuation to remove singular configurations.  The addition of redundancy to the parallel manipulator can increase the usable workspace of the robot.