Kernel PCA and approximate pre-images to extract the closest ultrasonic arc from the scanning of indoor specular environments

Processing of Lidar and IMU Data for Target Detection and Odometry of a Mobile Robot

“…The discriminating variable for the grouping of the elements depends on the problem proposed; in the case of point cloud data, it is the distance between the points. In this work, the clustering has been carried out by using the "kmeans" function [15][16][17]30]. The "kmeans" function takes as input the coordinates of the point cloud on which clustering will be performed and the desired number of partitions, while it outputs the indices of the points indicative of the assignment of points to one of Articles 9 the partitions, the coordinates of the centroids, and the distance of all points from each centroid.…”

“…The algorithm presented here, on the other hand, uses the IMU sensor and the LiDAR sensor, with the latter being capable of providing the frames containing point cloud data with a defined sampling rate, thus characterizing the neighboring environment. This makes it more versatile in eliminating critical issues, with a similar methodology to that used for an array of ultrasonic sensors for scanning external environments, determining useful characteristics for the reconstruction of simple external reference surfaces [15][16][17]. The algorithm makes use of a linear Kalman filter, which evaluates the dynamic state of a system starting from measurements affected by noise.…”