Date of Award
Spring 1993
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Electrical & Computer Engineering
Program/Concentration
Control Engineering
Committee Director
Oscar R. Gonzalez
Committee Member
John R. James
Committee Member
Derya A. Jacobs
Committee Member
Griffith J. McRee
Call Number for Print
Special Collections; LD4331.E55D64
Abstract
The most important property of an automated mechanical system is reliable execution of a specified task. Automated mechanical systems work best in precisely structured environments where all operations are repeatable. Typically, these systems are difficult to modify and expensive to install. Sensors can be used to detect disturbances, thus reducing hardware cost by offsetting the need for precise rigid construction components. Sensors, however, increase the complexity and cost of the control software. Efficient methods of reacting to sensor information during automated system operation are required to reduce software complexity and expense.
This thesis analyzes a robotic construction task and describes a new rule based system which could be used to coordinate the construction. Within NASA Langley's Automated Structures Assembly Laboratory a tetrahedral truss structure, representative of expected space structures, is constructed using an industrial robot equipped with a special purpose end-effector. The completed structure is eight meters in diameter and composed of 102 truss members. The robot is used to extract truss members from a storage canister and move the truss members to a position where they may be installed into the structure by the end-effector. Currently the robot motion is based on stored paths. These paths were developed in a trial and error process within the NASA facility. Successful construction relies on the repeatability of these paths which are specific to a particular robot and structure. An automated path planner is being developed to overcome these limitations.
This system uses a geometric model of the robot and truss structure to determine a suitable path. Thus, different geometry's may be constructed by changing the graphic model. Unfortunately, differences between the graphic model and the actual hardware prevent successful execution of the paths in an open loop fashion. Sensor information must be used to augment the path from the path planner. Currently the executive program supervising the construction proceeds in a serial fashion. Integration of automated path planning and sensor information will require an event driven system which can respond to asynchronous sensor information. In this thesis, a new rule based system will be proposed and a demonstration system developed that is capable of directing the robot during the construction operation. In addition, this thesis will detail the development of a force torque control strategy for a six degree of freedom robot. Force torque information from a wrist mounted sensor will be used to provide active compliance. Compliance is required during contact operations to minimize contact forces and prevent mechanical jamming. Two Cartesian force torque schemes, inverse kinematics and resolved rate, are analyzed in depth. The resolved rate scheme is shown to be more efficient when the entire control loop is analyzed. This is a new result which may be specific to robots with the kinematic configuration of the American Cimflex Merlin robot used within the Automated Structures Assembly Laboratory.
The thesis begins with a description of the assembly system, followed by development of a force velocity control strategy for the robot, and then the design of a rule based system to switch between the different control strategies used during construction. Key elements to the rule base design are discussed, including problem partitioning, the ability to respond to asynchronous events, and sensor information verification.
Rights
In Copyright. URI: http://rightsstatements.org/vocab/InC/1.0/ This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).
DOI
10.25777/9y85-q518
Recommended Citation
Doggett, William R..
"Robot Force Velocity Control Using Expert System for Control Mode Changes"
(1993). Master of Science (MS), Thesis, Electrical & Computer Engineering, Old Dominion University, DOI: 10.25777/9y85-q518
https://digitalcommons.odu.edu/ece_etds/594
Included in
Controls and Control Theory Commons, Dynamics and Dynamical Systems Commons, Robotics Commons