University of South Florida
Explainable and Cooperative Autonomy Across Networks of Distributed Systems
Abstract
dc:descriptionLarge networks of complex systems-of-systems are commonplace and evermore present in both mundane and extraordinary facets of human existence. From the exponential growth of connectivity via the internet and other information networks, to the miniaturization of computers and sensors, to cross-domain sensor and communication networks, these networks of distributed systems-of-systems (NDSS) present incredible benefits and challenges. Autonomy is perhaps the most important and most difficult to achieve enabling technology for efficient performance of the NDSS. Giving each individual agent in a network the ability to manage its internal state in dynamic operating environments and in pursuit of multiple complex and possibly conflicting individual and network-level goals is key to efficient, robust, and adaptable performance. When humans are an integral part of the system through oversight, operation, physical interaction, or dependency, the safety, security, understandability, and verifiability of an autonomous agent's actions are paramount. These attributes and constraints describe a class of problems with members including autonomous ground vehicles operating on public motorways, autonomous air traffic control systems, constellations of Earth-observation satellites, and human-rated interplanetary spacecraft. In this work, the problem of how to enable a high degree of on-board system autonomy for individual agents within the human-centered NDSS is explored. Foundations for agent-based autonomy are discussed in the context of understanding and managing complexity. A Hierarchical Ensembles of Autonomous Decision Systems (HEADS) framework is proposed. The HEADS framework is designed to enable cooperative, explainable, and robust multiple objective decision making by collections of fuzzy logic expert systems based on linguistic variables. The HEADS framework is applied to a simulated constellation of Earth satellites tasked with managing sensor activation and maintaining battery status and subject to the dynamic space environment and fault conditions. The initial system design shows acceptable behavior and robustness against fault conditions. Performance and operational analysis provide the basis for manually tuning the system which increases performance by 48%. An additional 6.8% performance increase is seen after automated genetic algorithm based parameter tuning. The HEADS framework is shown to provide robust performance for the multiple objective dynamic decision making problem for a large network of distributed systems of systems.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy (Ph.D.)
- Level thesis:degree_level
- Dissertation
- Discipline thesis:degree_discipline
- Electrical Engineering
- Grantor
- University of South Florida
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Jorgensen, Peter Joseph
- Contributors dc:contributor
-
- Robert H. Bishop, Ph.D.
- Kwang-Cheng Chen, Ph.D.
- John Licato, Ph.D.
- Mia Naeini, Ph.D.
- Richard Will, Ph.D.
Subjects
dc:subject × 7Identifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.usf.edu/etd/10302