An application of data fusion techniques for indoor flight of drones

In order to measure the position of an indoor unmanned aerial vehicle (UAV), A combined position estimate system is designed and tested in this thesis. Two sensors, an optical flow sensor (Px4flow) and a ultrasonic sensor (Marvelmind), are employed to provide the raw measurement data. The extended Kalman filter (EKF) is used as the main data fusion method and some algorithms are built based on this, including the attitude estimation, angular rate compensation and the position estimation. A single board computer (Raspberry Pi) is employed and plays a role as the central controller. Matlab, C language and Python are used to produce all the programs. Some simulation and on-board tests are performed and will be discussed in the thesis as well. By analysing the test result it can be proved that this system is capable of providing an accurate position estimate (with standard deviation of 0.0275 meter) at a high update frequency (100 Hz).