Improvement of aircraft accident investigation through expert systems

Milosovski, G 2008, Improvement of aircraft accident investigation through expert systems, Doctor of Philosophy (PhD), Aerospace, Mechanical and Manufacturing Engineering, RMIT University.

Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Milosovski.pdf Milosovski.pdf application/pdf 4.52MB
TheToolGP1020.xls TheToolGP1020.xls Click to show the corresponding preview/stream application/octet-stream 3.98MB
Title Improvement of aircraft accident investigation through expert systems
Author(s) Milosovski, G
Year 2008
Abstract Research was conducted on the ‘Improvement of Aircraft Accident Investigation through Expert Systems’. This research project aimed to analyse aircraft accident investigation and to review some of the off-the-shelf tools that support these investigations. An analysis of the investigation process and tools provided a possible avenue for improving aircraft accident investigation processes and for implementing mitigation measures to enhance air traffic safety. The work concludes that an intuitive and interdisciplinary approach must be vital elements of any contemporary method used for establishing a set of priorities for further improvement of aircraft accident investigation. Hence, the interpolations methods, including an analysis of accident statistics and Delphi enquiry, are appropriate tools in analysing and drawing conclusions for further improving the investigation. A Delphi enquiry, which is formalised regarding the procedure for carrying out research activities, has provided a comprehensive analysis of the investigation procedure including investigation techniques. A team of experts created for this purpose has conducted a qualitative and quantitative analysis of the factors having an impact on investigation outcomes. The Delphi study has indicated that there is great potential for further improvement of aircraft accident investigation. It has pointed out areas of the investigation process where significant improvements could be achieved. Moreover, the Delphi method has shown that investigations could be significantly improved with the application of a global expert system as a tool for storing and analysing the forensic data of aircraft accidents worldwide. As a consequence, the computer program GP1020 has been created in order to demonstrate how expert systems could contribute to facilitating and enhancing investigation results. The outcome of GP1020 is a novel investigation tool in the form of a data mining method designed to improve forensic data use by the aircraft accident investigators. The GP1020 program interface asks the user a tree-based set of questions related to conditions of wreckage, accident site and other circumstances relevant to accidents/incidents. Given enough information, the program is capable of narrowing down all known possibilities to indicate the most probable causes of the accident/incident. This work provides the answers to several questions relating to the improvement of investigation outcomes. First of all, the research indicates the factors that could contribute to obtaining better results of an aircraft accident investigation. Next, this research shows the facilities and procedure resources that have an impact on outcomes of aircraft investigations. Additionally, this study explains that a global expert system is a useful tool in aircraft accident investigation. This work also demonstrates how an expert system could enhance the investigation results. Moreover, in view of the growth in air traffic, this research points out the possible measures for reducing the number of aircraft accidents and their severe consequences.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Aerospace, Mechanical and Manufacturing Engineering
Keyword(s) Aircraft Accident
Delphi Technique
Aircraft Safety
Version Filter Type
Access Statistics: 615 Abstract Views, 12818 File Downloads  -  Detailed Statistics
Created: Tue, 30 Nov 2010, 13:06:57 EST by Guy Aron
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us