This MSc assignment goal is to enhance the level of autonomy of the aerial robots developed at RAM-UT. The available robot at UT should be able to effectively perform handovers to humans in both indoor and outdoor environments in an autonomous manner. To accomplish this task, the aerial vehicle must be provided with an accurate self-localization system in GPS-denied environments. This would grant the system with capabilities to determine its position and pose and required information for obstacle avoidance or object tracking.
This work analyses the main odometry techniques and completes an evaluation of the latest open-source VI-SLAM algorithms, with the final goal of performing localization in micro aerial vehicles (MAVs). The best VI-SLAM implementations will be tested against the well known EuRoC dataset and against a custom dataset which is more specific for our use case. The results are analysed to find the most convenient option to be used for localization RAM-UT MAV.