Real-Time Scan Matching for Indoor Mapping with a Drone

This paper aims to develop a method to mainly determine the 2D planar position of an aerial robot using just one horizontal layer from a laser scanner sensor onboard. Our proposed approach employs a modified scan matching method based on an Iterative Closest Point (ICP) algorithm. The novel aspect of the approach is using an elliptical weighting for the corresponding pairs within the ICP algorithm based on their direction from the point source, hence giving more weight to the points at the side boundaries. A numerical simulation experiment was conducted, and we showed that our proposed elliptic weighting is most beneficial for missions requiring high pitch angles. Compared to traditional ICP, when applying the proposed elliptical weighting at high pitch angles, it can reduce the error in estimating the robot’s position from 10% to 5%. The developed method was then successfully employed while using the horizontal layer from 3D LiDAR scans to estimate the transformation matrices and generate a reconstructed map for a confined space within a simulated cement plant.