An Evidence Theoretic Approach to Design of Reliable Low-Cost UAVs
Abstract
Small unmanned aerial vehicles (SUAVs) are plagued by alarmingly high failure rates. Because these systems are small and built at lower cost than full-scale aircraft, high quality components and redundant systems are often eschewed to keep production costs low. This thesis proposes a process to "design in" reliability in a cost-effective way. Fault Tree Analysis is used to evaluate a system's (un)reliability and Dempster-Shafer Theory (Evidence Theory) is used to deal with imprecise failure data. Three unique sensitivity analyses highlight the most cost-effective improvement for the system by either spending money to research a component and reduce uncertainty, swap a component for a higher quality alternative, or add redundancy to an existing component. A MATLAB (registered trademark) toolbox has been developed to assist in practical design applications. Finally, a case study illustrates the proposed methods by improving the reliability of a new SUAV design: Virginia Tech's SPAARO UAV.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 28, 2009
- Accession Number
- ADA522686
Entities
People
- Justin F. Murtha
Organizations
- Virginia Tech