Co-operative conflict resolution in autonomous aircraft operations using a multi-agent approach

Ruiz, Miguel Angel Vilaplana (2002) Co-operative conflict resolution in autonomous aircraft operations using a multi-agent approach. PhD thesis, University of Glasgow.

Abstract

The future Air Traffic Management is anticipated to feature Autonomous Aircraft Operations in specifically designated airspace. Aircraft flying in this airspace will be referred to as Autonomous Aircraft and will need to meet certain airborne equipment requirements. Autonomous Aircraft will be free to fly operator-preferred routes and their flight crews will have the task of maintaining safe separation from other Autonomous Aircraft without the intervention of Air Traffic Control. A conflict resolution methodology will be in place to enable flight crews to prevent violations of the established separation minima. To apply the conflict resolution methodology, the flight crews may need the assistance of on-board decision-support tools. Autonomous Aircraft Operations have the potential to improve the safety and efficiency of flight operations in regions of airspace with no radar-based Air Traffic Control coverage. This thesis investigates the potential application of Distributed Artificial Intelligence concepts and techniques to develop co-operative conflict resolution methodologies for Autonomous Aircraft Operations. In this context, the term "co-operative" is used to describe conflict resolution methodologies by which conflicting Autonomous Aircraft safely co-ordinate their resolution actions so that the resolution costs are shared equitably amongst all the Autonomous Aircraft involved. A new approach to conflict resolution in Autonomous Aircraft Operations is proposed based on the sub-field of Distributed Artificial Intelligence concerned with the study of multi-agent systems. This new approach designates Autonomous Aircraft as intelligent agents and considers conflict resolution as a co-operative activity in the framework of a multi-agent system. The means necessary for co-operation in a multi-agent system are provided by a cooperation mechanism. Following a review of multi-agent systems research literature, two main types of co-operation mechanisms have been identified: behaviouristic and reflective. The former type refers to co-operation mechanisms that allow the agents to be seen as acting co-operatively by an external observer, regardless of whether or not the agents are knowingly co-operating. The latter type refers to co-operation mechanisms that allow agents to engage knowingly in co-operative activity, regardless of whether or not they are seen to act co-operatively by an external observer. To illustrate the capabilities of the proposed multi-agent approach, two examples of cooperation mechanisms that could be implemented as co-operative conflict resolution methodologies in Autonomous Aircraft Operations are presented. The first co-operation mechanism is of the behaviouristic type and has been developed specifically for an operational environment where Autonomous Aircraft can only exchange information with one another through Automatic Dependent Surveillance-Broadcast. The second co-operation mechanism is of the reflective type and has been developed specifically for an operational environment where Autonomous Aircraft can also exchange information with one another on a one-to-one basis using a point-to-point digital data-link. The two co-operation mechanisms include the necessary algorithms, protocols and procedures to design on-board decision support tools that could aid flight crews in resolving conflicts co-operatively. The performances of the co-operation mechanisms in two-dimensional conflict scenarios, involving up to three aircraft, are analysed and compared.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: Systems science, air traffic management, autonomous aircraft operations.
• Subjects: T Technology > T Technology (General)
• Colleges/Schools: College of Science and Engineering > School of Engineering
• Supervisor's Name: Goodchild, Dr. Colin
• Date of Award: 2002
• Depositing User: Enlighten Team
• Unique ID: glathesis:2002-71064
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 10 May 2019 10:49
• Last Modified: 18 May 2021 11:06
• URI: https://theses.gla.ac.uk/id/eprint/71064