This is a version that runs CUSUM on real-time simulated data, displaying the signal thanks to graphview.

Paper relative to the project code can be found here : paper

Abstract

Mapping, surveillance, video, photography, shipping or science and research, the areas of application of micro UAVs ( Unmanned Aerial Vehicle) are numerous. In order to ensure the smooth running of the drone’s mission, it is often crucial to precisely control its trajectory and to be able to detect any change in its 3D orientation. Inertial sensors provides angular information along different axes. The objective is to minimise the average detection delay cost of change in the statistical property (the mean) of one of these signals. To do so, a quickest change detection algorithm known as CUSUM has been chosen. The purpose of this project would be to find a strategy to adapt and implement Bruggemann’s version of Cusum algorithm (written in Matlab programming language) in the mobile SDK(remote controller) of the DJI matrice 600 or DJI Phantom in order to process information from the drone, in real-time.