This app was developed for my master's thesis at the University of Oslo (UiO). The thesis was supervised by Prof. Paulo Ferreira and written in English.
flyDetect is an app capable of reliably detecting flight when onboard most airliners. It does this using only the built-in sensors of the device it's running on - the accelerometer and barometer.
Out of 31 flights, flyDetect successfully detected takeoff in 100% of them and landing in 95.7%. Because of the simplicity of the algorithms responsible for detection and the use of low-power sensors, flyDetect has an impressively low power consumption. With the algorithm in the "not flying" state, flyDetect consumes an estimated 0.40 mA (approximately 0.01% battery drained per hour on the test device).
The dataset used to develop and evaluate flyDetect is not publicly available. See Dataset availability for more information.
- Read the paper - a condensed version of the thesis, published in Sensors.
- Read the thesis - the full master's thesis.