Thesis
ECS: an adjustable autonomy framework for the Soar agent architecture
Washington State University
Master of Science (MS), Washington State University
2011
Handle:
https://hdl.handle.net/2376/102982
Abstract
We are at a point in time where we can better leverage the usability and flexibility of computers with more complex requirements of the user. Because of this, machines are now responsible for a wide variety of tasks without supervision. However, some cases will arise where the machine's internal knowledge or capabilities may be lacking, sub-optimal, or unsafe in a particular situation. When this occurs, we look to other means to ensure correct behavior. By providing the user with a means to monitor and regulate the reasoning process of a machine, we manage machine-induced failures and introduce another safeguard into the system. We present an overview of (adjustable autonomy). We then describe the Soar agent architecture which utilizes a unique behavioral model and goal structure for simulating real-world problem solving. We will then describe a unique agent validation framework implemented within Soar from which we borrow several core ideas. We finally describe our implementation of adjustable autonomy, as an extension to the existing self-assessment framework.
Metrics
2 File views/ downloads
8 Record Views
Details
- Title
- ECS
- Creators
- Matthew Henry
- Contributors
- Scott Wallace (Degree Supervisor)
- Awarding Institution
- Washington State University
- Academic Unit
- Electrical Engineering and Computer Science, School of
- Theses and Dissertations
- Master of Science (MS), Washington State University
- Publisher
- Washington State University; Pullman, Wash. :
- Identifiers
- 99900525174401842
- Language
- English
- Resource Type
- Thesis