M. S. Thesis Abstract
An Experimental Investigation of a Robust Force Control Algorithm in the Presence of a Compliant Environment
The ability to change the dynamic characteristics of an environment by a robot is limited if only position information is known. To overcome this limitation, research emphasis is being placed on force and impedance controllers that can sense the location of items by "feeling" them. When disturbances and uncertainties are present, however, linear controllers have difficulty achieving the desired force response.
The purpose of this research was to experimentally implement a nonlinear deterministic robust controller and to evaluate it relative to linear state feedback control for force control in the presence of uncertainties. The robust controller consisted of a full state feedback control contribution and a robust control contribution. A test platform was constructed from two single degree-of-freedom (DOF) linear actuators; the controllers could be evaluated in the presence of friction, and disturbances could be added electronically through a general second order environment. The system emulated a robot acting upon an environment which might be found in a typical manufacturing application.
When the robust control contribution was added to full state feedback using a general set of poles, performance improvements were realized in force step response tests and force regulation experiments. In the step response test, the robust controller greatly improved the steady state accuracy of the system, and in some instances removed oscillatory behavior. During force regulation, the robust controller also showed improvements over the performance of the state feedback controller.