Date of Award

12-2017

Document Type

Dissertation - SCU Access Only

Publisher

Santa Clara : Santa Clara University, 2017.

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical Engineering

First Advisor

Christopher A. Kitts

Second Advisor

Ignacio A. Mas

Third Advisor

Sergio Zarantonello

Abstract

Robotic systems are increasingly being used for tasks that humans are incapable or unwilling to perform. For some automated tasks, groups of robots can provide better or more cost effective performance than a single robot alone. One area of research for multirobot collaboration is transportation and manipulation of large objects. Multirobot transportation of objects can be enhanced by using a hybrid force/position control architecture, a technique believed to have been used with mobile robot formations for the first time in this work and associated publications. The use of this hybrid architecture allows for position control of some degrees of freedom while controlling formation level forces associated with other degrees of freedom. For the specific case of object transportation, the size and relative location of a formation along an object can be maintained while the formation applies a specified net force and torque on the object. The hybrid architecture provides a natural and simple method of specifying formation level movements and interaction forces. Furthermore, using the hybrid architecture provides seamless transitions between pure position control, for approaching and disengaging the object, and hybrid force/position control, for pushing the object along a path or to a specified location. Additional benefits of the architecture include abstracting the robots as a virtual actuator in the position control process of the object being transported as well as the prevention of large and potentially dangerous interaction forces. The hybrid force/position control architecture is formally described in this dissertation and validated with both simulation and hardware experiments.

Download

Share

COinS