General Aviation (GA)
accidents constitute the majority of aviation related accidents. In the United
States, there have been over 7,000 GA accidents compared to 190 airline
accidents in the last 8 years. Flight data analysis has helped reduce the
accident rate in commercial aviation. Similarly, safety analysis based on
flight data can help GA be safer. The FAA mandates flight data recorders for
multi-engine and turbine powered aircraft, but nearly 80% of General Aviation
consists of single engine, of which only a small portion contain any form of
data recording device. GA aircraft flight data recorders are costly for
operating pilots. Low-cost flight recorders are few and rarely used in GA
safety analysis due to lack of accuracy compared to the certified on-board
equipment. In this thesis, I investigate the feasibility of using a low-cost Attitude
and Heading Reference System (AHRS) to detect hazardous states in GA aircraft. I
considered the case of roll angles and found that the low-cost device has
significant measurement errors. I developed models to correct the roll angle
error as well as methods to improve the detection of hazardous roll angles. I devised
a method to evaluate the time accuracy along with the angle accuracy and showed
that despite the errors, the low-cost device can provide partial hazardous
state detection information.