| description abstract | Road roughness, defined by the International Roughness Index (IRI), is a critical criterion for ride quality and comfort, meticulously monitored by road authorities to address maintenance needs. This paper introduces a new method to explore the suitability of bicycles as probe vehicles for measuring nonmotorized road roughness. For this purpose, polyvinylidene fluoride (PVDF) sensors are attached to the front wheel of a mountain bike to capture road roughness through tire–road interaction. To validate this approach, a study was conducted on a motorized dual-lane road, where each direction spanned 660 m, totaling 1,320 m, to verify the method’s accuracy in measuring IRI. Data from both the PVDF sensors and their specific locations were recorded simultaneously. The values obtained from a laser profilometer vehicle served as benchmark reference points for the PVDF sensor readings. Thirty-two features are extracted from the PVDF sensor data. The Support Vector Regression (SVR) algorithm is then used to estimate IRI values from these features. The mean absolute percentage error (MAPE) results of the data sets for the distances covered by 15, 30, and 50 full rotations of the bicycle’s front wheel, corresponding to 30, 60, and 100 m, respectively, are found to be 13.64%, 10.73%, and 5.34%. These results highlight the potential of this innovative approach as a reliable tool for determining road roughness on nonmotorized pathways. | |
