On the Way to a Real-Time On-Board Orthogonal SLAM for an Indoor UAV

Here we present the way to a real-time on-board SLAM (simultaneous localization and mapping) algorithm of a quadrotor using a laser range finder. Based on successfully implemented techniques for ground robots, we developed an orthogonal SLAM algorithm that merges a new scan into the global map without any iteration. This leads to a low requirement of computing power and computing time for the SLAM calculation. The algorithm delivers a 2D floor plan of the investigated area. All essential computations will be clone on a microcontroller which is mounted on an industrial quadrotor. Due to the fact that all calculations including the SLAM algorithm will run on-board the robot is able to act autonomously in an unknown indoor environment. To enable a robot to act autonomously several navigation controllers are needed. The basic ideas and the implementation are also part of this paper. Finally, it comprises the results of an autonomous indoor flight of the industrial quadrotor AR100B(R) of the AirRobot(R) Company equipped with a self constructed functional group.