Use this URL to cite or link to this record in EThOS:
Title: An intelligent agent based autonomous mission management system for UAVs
Author: Gunetti, Paolo
ISNI:       0000 0004 2716 358X
Awarding Body: University of Sheffield
Current Institution: University of Sheffield
Date of Award: 2012
Availability of Full Text:
Access from EThOS:
Access from Institution:
Unmanned Aerial Vehicles (UAVs) have been around for almost eight decades, but they evolved from the basic remotely-controlled form only during the last two. While UAV-related research encompasses several areas, the increase of autonomy is certainly one of the most important topics. Current-generation UAVs are typically able to autonomously execute a pre-planned mission, and there is a definite trend towards increased UAV autonomy. The main challenge related to UAV autonomy is the reaction to unforeseen events; the capability to execute pre-planned tasks does not take into account the fact that in a dynamic environment the plan might need to be changed. This Mission Management activity is usually tasked to human supervision. Within this thesis, the problem of UAV autonomy will be addressed by proposing a software architecture that allows the integration of various technologies to obtain a significant improvement in the autonomy level of UAV. The first step in this is the definition of a set of theoretical concepts that allow the computational description of three different types of information: user-generated mission objectives, knowledge regarding the external environment and complete flight plans. The second step is then to develop a software system that autonomously accomplishes the Mission Management task: combining mission objectives and environmental knowledge to generate a viable flight plan, then update it when situational awareness changes. The software system presented in this thesis is implemented using a combination of three separate Soar Intelligent Agents and traditional control techniques. Soar (State, Operator and Result) is the computational implementation of a general theory of cognition, allowing for the definition of complex software agents which can efficiently apply the rules defining their behaviour to large amounts of knowledge. Soar agents are used to provide high-level reasoning capability to the system, while low-level control functions are better performed by traditional algorithms. The software system is fully described and implemented, then tested in a simulation environment. Simulations are used to demonstrate the system's ability to automatically generate, execute and update (when necessary) an entire flight plan after being assigned a set of mission objectives.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available