| Résumé : |
(auteur) In recent years, Unmanned Aerial Vehicles (UAVs) have been used increasingly as mobile mapping platforms for kinematic applications in the field of geodesy. For this purpose a UAV is usually equipped with a mapping sensor, such as a camera or a laserscanner. A major step to make the collected data useful for surveying applications is the georeferencing, which associates the images or range measurements and the derived products (point clouds, orthofotos, 3D models) with a spatial reference. In this dissertation the development of a direct georeferencing system for real-time position and attitude determination of lightweight UAVs is presented. The term „direct" means that the georeferencing is based on an onboard multi-sensor system. Sensors, which are typically used in this context, are GPS (Global Positioning System) receivers, inertial sensors and magnetometers. For geodetic UAV-applications Micro-UAVs are usually applied, which have a weight limit of 5 kg. Therefore, weight, size and power supply constraints play an important role here. As a result of these limitations inertial sensors based on MEMS (Micro Electro Mechanical System) technology are mostly used. This technology enables the compliance with size and weight limits, but it also leads to significant drift effects in the position and attitude determination after a short period of time. To bound these drift effects and to be able to provide high accuracies (e.g. position: 5 cm, attitude: 0.5 deg) in the direct georeferencing of a Micro-UAV, a high availability of precise GPS carrier phase measurements is crucially important. As a consequence the following aspects will be addressed in this dissertation:
• GPS carrier phase measurements are ambiguous by an unknown number of integer cycles. These so called ambiguities have to be resolved after every loss of lock of the satellite signals to fully exploit the high accuracies of the carrier phase observables. Since the satellite signals are interrupted frequently during kinematic applications, procedures are developed, implemented and evaluated, which enable a fast ambiguity resolution and allow for a high availability of CPS carrier phase measurements under challenging GPS measurement conditions.
• With the aim to realize high accuracies and a high robustness, redundant information from several sensors is integrated in a sensible manner in sensorfusion algorithms In order to be able to deal with challenging GPS measurement conditions, the sensorfusion is realized at the level of GPS raw measurements in a tightly-coupled GPS/MEMS-IMU integration algorithm. In this way GPS carrier phase measurements can even be used, if less than four satellites are visible, which also increases the availability of a precise position estimation. The accuracy improvements of a tightly-coupled over a loosely-coupled integration during challenging GPS measurement conditions are investigated.
• The algorithms are implemented on a multi-sensor system, which has been developed for the direct georeferencing of lightweight UAVs in this dissertation. Results of flight tests and measurements with a portable test system demonstrate that the developed direct georeferencing system leads to position accuracies of less than 5 cm and attitude (roll, pitch, yaw) accuracies of less than 0.2 deg, if GPS carrier-phase measurements are available. Overall, this dissertation gives detailed insights into the development of algorithms and a multi-sensor system for the direct georeferencing of lightweight UAVs in real-time. The findings gained in this thesis are not only valid for the position and attitude determination of lightweight UAVs but also for other mobile platforms, such as cars, ships, airplanes or rail-borne vehicles. Therefore, this work makes an important contribution to a current trend in the field of engineering geodesy, where mapping, monitoring and also setting-out is more and more realized using mobile mapping systems. |
Entwicklung einer direkten Georeferenzierungseinheit zur Positions- und Orientierungbestimmung leichter UAVs in Eichzeit (2016)
