ROAD PROFILE IDENTIFICATION BY MEANS OF REGULARIZED LEAST SQUARE MINIMIZATION WITH DYNAMIC PROGRAMMING UTILIZING ACCELERATIONS OF A MOVING VEHICLE

Abstract

 This study aims to improve accuracy of the road profile identification utilizing accelerations of a moving vehicle. There exist road profile identification methods: one identifies the vehicle force by dynamic programming and convert it into road profile and the other identifies road profile directly by pseudo inverse matrix. This study proposes a new method combining advantages of those two approaches that identifies road profile directly by means of a regularized least square minimization with dynamic programming. Accuracy of the proposed method is investigated by numerical simulations and a field experiment. Comparative investigations between the proposed method and the method based on Kalman filter to identify the road profile are also carried out. Observations showed that the proposed regularized least square minimization with dynamic programming method could be a practically useful method in IRI estimations utilizing acceleration responses of vehicle bounce motion.