A PCA-based interpretation of ultrasonic data for accurately detecting a corridor wall

Abstract

This paper introduces a validated model for the ultrasonic detection by processing sensor data and it presents a new approach for evaluating the orientation of a regular wall. The method described is based on the Principal Component Analysis (PCA) which gives the opportunity to make some geometrical considerations on the effectiveness of the scanning results. Several experiments are conducted scanning a regular corridor wall and also propagating ultrasonic waves towards a wall opening. In every instance these tests permit to confirm the applicability of the new proposed strategy.