Use this URL to cite or link to this record in EThOS:
Title: Fusing knowledge and image sensor data for mission-critical control in unmanned aerial vehicles
Author: Patterson, Timothy John
Awarding Body: Ulster University
Current Institution: Ulster University
Date of Award: 2013
Availability of Full Text:
Access from EThOS:
Autonomous Unmanned Aerial Vehicles (UAVs) have the potential to significantly enhance current working practices for many applications including environmental monitoring, aerial surveillance and mountain search-and-rescue. Their ability to operate as an 'eye-in-the-sky', relaying real-time aerial imagery and other sensory data whilst removing humans 'from situations which may be considered dull dangerous or dirty ensures that the popularity and usage of such platforms will continue to increase. However, as with manned aircraft, the dependability and integrity of such platform may be influenced by the occurrence of endogenous and exogenous events ultimately resulting in safety-critical and , possibly, mission-critical situations. Such events will inevitably cause a range of errors and may ultimately result in the loss of human life, damage to property and the destruction of the UAV platform and its payload. With this in mind: the UK Civil Aviation Authority currently impose similar restrictions on small UAVs to those specified for model aircraft, thereby limiting their real-world usefulness. Before these restrictions can be relaxed and the potential of UAVs realised they must be able to demonstrate equivalent levels of safety to human pilots in safety-critical situations. As such, one objective is a forced landing scenario in which a UAV encounters an error and is required to locate and land in a Safe Landing Zone (SLZ). Within this thesis we present a programme of work aimed at researching and developing an autonomous method of UAV SLZ detection using colour aerial imagery and knowledge in the form of Ordnance Survey (OS) data. The key outputs of this thesis are as follows: 1) a method of position estimation which extends previous approaches by utilising OS data as opposed to geo-referenced aerial imagery 2) an autonomous method of SLZ detection which exploits colour aerial imagery and knowledge in the form of OS and training data 3) an approach which combines multi-resolution aerial images of the same scene with OS and training data to compute updated class parameters which are subsequently utilised to perform terrain classification 4) a method of UAV decision making within a time-constrained: safety-critical situation which is based upon constructing models of execution times. We evaluate the developed algorithms using representative aerial imagery which was captured during manned flight by piloted aircraft and present results demonstrating practicable potential in the proposed approaches.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available