| Abstrakt: |
ABSTRACTIn our previous work, a new technique for dynamically distributing the load imposed by a rigid object between two manipulators which mutually hold the object was proposed. The technique was applied given two different choices of a designer specified matrix, and in each case it resulted in an optimal solution for the contact forces and torque imparted to the object by the manipulators. Each contact force solution contains an object motion inducing component and an internal stress and torsion inducing component. The internal component is a linear function of a vector of “corrective” contact forces. This article addresses the problem of determining the corrective contact forces such that the internal component of the contact force solution contributes to contact force magnitude limit avoidance. The basic premise to our approach is that there is a maximum magnitude limit associated with each of the contact forces, and that a magnitude limit avoidance algorithm should be activated only when it is detected, by sensing, that one or more of the contact forces are in close proximity to their respective magnitude limits. Two algorithms for calculating the corrective contact forces are proposed and compared to determine their relative advantages. The article also presents several new insights into the modeling and optimization technique presented in our earlier work. |
