Abstract
This work focuses on autonomous operation of rotary-wing Micro Aerial Vehicles (MAVs) in previously unexplored, GPS-denied environments. This type of operation requires the MAV to simultaneously estimate a map of its surroundings and its own position within that map. This work presents a refined algorithm for performing Simultaneous Localization and Mapping (SLAM) which does not rely on revisiting previously mapped areas. The algorithm is first benchmarked against previously published algorithms. Different scenarios, both indoors and outdoors, are then presented, showing the high amount of detail captured by the proposed method. Scenarios were mapped using three different platform types, including a conventional main-tail rotor helicopter, showing that the algorithm is platform-independent and does not require the rotorcraft's dynamic model, which can be complex. Experimental validation of the SLAM method shows an accuracy of approximately 0.1% of the traveled course length. The algorithm is capable of operating in real time on small form factor computers.