Optimized Real Time 3-D Feature Map Generation with Unmanned Aerial Systems

Abstract

The advancement in unmanned aerial vehicle (UAV) automation has greatly increased the application of autonomous UAVs in both civilian and military tasks. Powerful yet light weight onboard computers are now available to further reduce the size of UAVs to a great extent and thus shrinking them to a size suitable for indoor missions. The main challenge with indoor navigation is usually the lack of GPS reception. Thus, simultaneous localization and mapping algorithms (SLAM) must utilize proximity sensors such as LiDAR, SONAR, and Stereo Cameras. In order to keep the size and weight of the UAV to a minimum, extracting more information from a limited number of sensors is vital. Hence, this research aims at the improvement, implementation and testing of localization and 3D mapping algorithms using a single rotating 2D LiDAR.