Road surface estimation: Exploring road surface estimation as a part of condition monitoring

Abstract

This thesis explored if it is possible to estimate road surface by measuring the acceleration of the vehicle’s suspension, sprung and unsprung mass. Being able to estimate the road surface can be advantageous in condition monitoring as traveling on different road surfaces affects the wear of the vehicle in different ways. The benefits and drawbacks of the equipment used were assessed in order to establish which is the most efficient at road surface detection. To estimate road surface a car was equipped with seven strategically placed accelerometers and data acquisition devices to gather data. The car was then driven upon asphalt and gravel roads of varying quality. Data was also gathered from the cars CAN-bus. The acquired data was analysed using Peter D. Welch’s power spectral density estimate to identify differences between the two surfaces. Magnitude squared coherence was used to determine the relation between the accelerometers. While the study suggests that road surface detection by the means of measuring acceleration seems achievable. The results need to be verified with larger sets of data. The thesis concluded that not all accelerometers mounted were needed as many showed the same results. Measuring the acceleration of the vehicles suspension acceleration was deemed to be the most accurate way to estimate road surface.