PCA-Based Classification for 3D Ultrasonic Reflectors

In this work a classification system for basic ultrasonic reflectors (plane, corner and edge) in three-dimensional (3D) environments is presented. The classification system is based on the principal component analysis (PCA) technique and the times-of-flight (TOF) are used as classification parameter. The system provides a sensorial structure to simultaneously obtain up to 16 TOF in every emission/scanning process. With the set of obtained TOF, it is possible to identify the ultrasonic reflector, as well as its position and range in 3D environments. The results achieved by the classification system are satisfactory, since the three reflector types are identified and located in 3D environments.