Akademik Veri Yönetim Sistemi
Tezin Türü: Yüksek Lisans
Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Elektrik ve Elektronik Mühendisliği Bölümü, Türkiye
Tezin Onay Tarihi: 2018
Öğrenci: TALHA İNCE
Danışman: AFŞAR SARANLI
Özet:An Inertial Navigation System (INS) is a combination of hardware (accelerometers and gyroscopes) and algorithms to calculate the position, orientation and velocity of a mobile platform. Because of the need to integrate the measurements over time, INS is subjected to cumulative error characteristics, hence cannot provide an accurate navigation solution over long durations. Global Positioning System (GPS) is often used for long time-long distance problems aiding INS. GPS relies on external signals received from satellite networks which gives bounded error. Unfortunately, these signals can be jammed, spoofed or may not be available at indoor applications. Hence, alternatives to the GPS sensor are required. Such a recent alternative is the camera sensor, resulting in Visual Aided Inertial Navigation System (VINS). VINS aims to provide accurate navigation solution using a fusion of imaging and inertial sensor data. The main goal of this thesis is to analyze and increase the performance of VINS sensor fusion algorithms on an unmanned aerial vehicle (UAV). For this purpose, a realistic simulation environment is implemented and different VINS methods are comparatively studied. Extensive simulation studies are conducted to characterize the performance of map-based and mapless VINS methods and to study the effects of performance critical parameters. Finally, a modified Gaussian Mixture Filter and feature region selection method are proposed for increasing the performance of VINS.