Simultaneous Localization and Mapping Algorithm for Unmanned Aerial Vehicles Based on Improved ICP

This paper addresses the issue of reduced localization and mapping accuracy in unmanned aerial vehicles (UAVs) equipped with LiDAR due to UAV motion and airframe vibrations. We propose an improved Iterative Closest Point (ICP)-based algorithm for UAV localization and mapping. Firstly, the voxel filtering algorithm is used to preprocess the point cloud data containing a large number of noise points generated by the UAV vibration. The original point cloud data is denoised and compressed while retaining the basic characteristics of the point cloud data. Secondly, the point cloud data obtained during the movement of the UAV will produce motion distortion. In this paper, the IMU data is coupled with the point cloud data to reduce the degree of point cloud distortion. Finally, an improved ICP algorithm is designed. The Euclidean distance between the calculated points in the traditional ICP algorithm is improved to the distance between the calculated points and the tangent plane, which improves the accuracy and speed of point cloud registration. The KITTI dataset is used to test the improved laser SLAM(Simultaneous Localization and Mapping) algorithm. The experimental results show that the average error and root mean square error of the improved algorithm are reduced by 31.57% and 27.25 % respectively compared with the original algorithm.