Graduation Semester and Year
Summer 2024
Language
English
Document Type
Thesis
Degree Name
Master of Science in Aerospace Engineering
Department
Mechanical and Aerospace Engineering
First Advisor
Animesh Chakravarthy
Second Advisor
Shuo Linda Wang
Third Advisor
Alan P. Bowling
Abstract
This thesis investigates a multi-agent containment problem where a fast evader, moving at a constant speed and heading, attempts to escape a circular containment region. A team of slow pursuers possess a nonzero capture radius and seek to capture the evader before escape. The research addresses two primary scenarios: one involving the team of pursuers, with no communication network, being constrained to the edge of the containment region, and the second involving a team of pursuers, with a communication network, not being constrained to the edge of the containment region. Both scenarios study the influence of malicious agents in the team of pursuers and the efficiency of a proposed boost countermeasure. In the first scenario, the study uses a constrained unicycle model with no communication network and explores various capture conditions, focusing on the worst-case initial position for the pursuer, defined as the position that allows for the largest angular range away from the evader's exit aim point. A parametric analysis examines the effects of speed ratio, capture radius, and the evader's initial location. A reachability analysis identifies viable escape headings and reachable regions for the evader. In the second scenario, a double integrator dynamic model incorporates a communication network and consensus laws. A directed tree communication network among a team of four pursuers is used for most example simulations. The consensus law allows for flight formation, and leader-following capabilities. Both scenarios consider situations where some pursuers are compromised and become malicious, initiating capture with a response delay. A boost countermeasure is proposed, and its effectiveness is evaluated.
The study aims to provide a comprehensive understanding of containment strategies in multiplayer pursuer-evader scenarios, addressing the challenges posed by malicious agents and enhancing strategic responses through both constrained and less-constrained models. This comprehensive study lays the groundwork for future research on more complex evader inputs, larger teams of pursuers, and more dynamic evader models, enhancing our understanding of containment strategies in multi-agent systems.
Keywords
Containment, Pursuit-evasion, Consensus, Multi-agent, Flight formation, Malicious agents.
Disciplines
Controls and Control Theory | Navigation, Guidance, Control and Dynamics
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Recommended Citation
Mora, Braulio, "Study of a Multi-Agent Containment Problem Under Communication Constraints and Defense From Malicious Agents" (2024). Mechanical and Aerospace Engineering Theses. 678.
https://mavmatrix.uta.edu/mechaerospace_theses/678