Attitude Estimation and Control of VTOL UAVs

The theoretical challenge involved in the operation of VTOL UAVs is often divided into two main problems. The first problem involves the development of an estimation scheme which can accurately recover the orientation, or angular position, of the aircraft. The second problem involves the development of algorithms which can be used to reliably control the orientation and/or the position of the vehicle. These two problems are the primary focus of this thesis. We first consider the problem of attitude estimation. To solve this problem we use vector measurements, and in many cases, a gyroscope (to measure the angular velocity of the system) in order to develop the estimation scheme. In the case where an accelerometer is used to provide a measurement of the apparent acceleration, we consider a special class of attitude observer, known as a \emph{velocity-aided} attitude observer, which additionally use the system linear velocity to improves the estimation performance when the system is subject to high linear accelerations. Secondly, we develop a number of algorithms which can be used to control the orientation and/or the position of the system. Two adaptive position tracking control laws are proposed which are able to compensate for exogenous disturbance forces. However, this control strategy (like other existing position control strategies) assumes that the system orientation is directly measured, where in reality only an estimate of the system orientation is available, which is obtained using some attitude estimation scheme. Therefore, we also propose an attitude stabilization control law, and two position control laws which do not assume that the system orientation is directly measured. To develop these control laws, we use vector measurements (that would normally be used by the attitude observer) directly in the control algorithms, which eliminates the requirement for an attitude observer. We also consider a special type of the vector-measurement-based position control laws which uses the accelerometer to measure the body-referenced apparent acceleration (rather than assuming only the gravity vector is measured). Therefore, this proposed control strategy may be better suited for VTOL UAVs, which are likely to be subjected to linear accelerations