A real-time on-board orthogonal SLAM for an indoor UAV

Abstract

Over the last years we developed a real-time on-board orthogonal SLAM (simultaneous localization and mapping) algorithm for an indoor UAV based on successfully implemented techniques for ground robots. The algorithm delivers a 2D floor plan of the investigated area. The robot is able to act with full autonomy in an unknown indoor environment because all essential computations are done on-board. The focus of this paper is on two key features in the topic of SLAM algorithms. The first one is the computation of the robot's movement and especially of the robot's rotation between two scans and the measurement of the robots orientation in an indoor environment in general. In this paper we present a very simple method based on angle histograms. The second one is the loop-closing problem. We present some results that will show that the loop-closing is possible with our very simple approach of SLAM algorithm. Finally, it comprises the results of an autonomous indoor flight of the industrial quadrotor AR100B® of the AirRobot® company equipped with a self-constructed functional group. In contrast to our former paper [1], all essential computations including the SLAM algorithm are done in real-time and on-board. © Springer-Verlag Berlin Heidelberg 2012.