A Road Surface Distinction Method Based on Integral of Reflected Wave Using Ultrasonic Sensors
Hidemitsu Arimura, Masahiro Yamamoto, Shenglin Mu, Shota Nakashima · 2020
In this paper, a new road surface distinction method using ultrasonic sensors is proposed. The ultrasonic sensor introduced for constructing the measurement device, combining with microcomputer and Bluetooth module. The measurement device, which can send ultrasonic waves to road surfaces and receive reflected wave, is designed to be fixed on a wheel chair for road surface distinction. The integration of the reflected waveform in time is called voltage-time integral. In previous study, it was confirmed that the strength of reflected wave from road surfaces depends on the distance between the measurement device and road surface, the incidence of the ultrasonic wave and the roughness of the road surfaces. In this paper, four types of road surfaces were distinguished: floor, mat, artificial turf, and gravel. A classifier was built with support vector machine (SVM) using the average and standard deviation of the voltage-time integral as features. The sample size was varied and the accuracy score for each was tested. Hence, the effectiveness of the road surface distinguish method using the average and standard deviation of the voltage-time integral as parameters was confirmed.