Inertial navigation algorithms based on symmetry-preserving theory

Nowadays, the usage of low-cost sensors on Unmanned Aerial Vehicles (UAV) has become more popular due to the increase in small UAV platforms. Some of them are used for research on new algorithms in all the different parts of flight dynamics research like flight control, navigation, data fusion, etc. Furthermore, these low-cost and small sensors are used on a huge range of new applications such as filming, photography and agricultural uses. For all of these purposes, the UAVs rely on a host of sensors to position, navigate and compute all the necessary data for the application they are designed for. The key part in that process is to obtain accurate and reliable state estimations from available measurements. Taking into account the comparably weak performance of employed low-cost sensors, algorithms to estimate the measured and non-measured variables are of primary importance. In this thesis, three Inertial Navigation Algorithms for a quadrotor UAV are implemented and compared to show its essential advantage and drawbacks. The first two algorithms are both designed based on Symmetry-preserving theory and differ from each other by the type of estimation: a non-linear observer and an Extended Kalman Filter (EKF), whereas the third one is an EKF without the Symmetry-preserving theory behind it. Besides numeric simulations, these algorithms are applied to real-time data to be able to evaluate the properties of each algorithm.