Intervening Factors in Pavement Roughness Assessment with Smartphones: Quantifying the Effects and Proposing Mitigation

Abstract

Researchers have linked pavements roughness to acceleration signals provided by smartphones, due to their simple handling and low cost, which might facilitate continuous data collection that is important for pavement management systems (PMS). However, the influence of several factors that may impair the quality of the smartphone data collection have not been clarified yet, which decreases the system reliability for real PMS applications. This paper presents a comparison between vertical accelerations provided by smartphones embedded in vehicles and international roughness index (IRI) values obtained from both rod and level and inertial laser profilometer surveys. The analysis considered the influence of variation of the speed, smartphone model, type of smartphone mount, number of measurement trips, length of segments, and vehicle engine vibrations in the correlation between IRI and root-mean-square of vertical accelerations (RMSVA). The results indicate that generic smartphone applications with IRI prediction models that do not consider the particularities of a calibration process and signal processing can lead to serious errors in the pavement roughness assessment.