Procedimentos para estimar a irregularidade longitudinal do pavimento por meio de veículos calibrados, utilizando-se dados de aceleração vertical obtidos de smartphones

• Autor(a) principal: Magalhães, Fabrício Helder Mareco
• Data de Publicação: 2019
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Federal do Ceará (UFC)
• Texto Completo: http://www.repositorio.ufc.br/handle/riufc/52918
• topic: Transportes; Smartphones; Pavimentos; Procedures

Currently, recent national and international surveys have been developed relating the signals provided by the accelerometers, installed in smartphones, to the International Roughness Index - IRI. Although these researches demonstrate that smartphones can be a viable tool for the evaluation of longitudinal irregularity, there are still doubts about the procedures that have been carried out to verify the irregularity, mainly in regard to obtaining the mathematical models of calibration of the vehicles. In this study, two experimental sections of road pavement were selected in which their respective International Roughness Index values were calculated by the Level and Rod method and the vertical acceleration signals data, later converted to RMSVA (Root Mean Square Vertical Acceleration) by means of a smartphone attached to the HB 20 and SUV ix35 vehicles with operating speeds ranging from 20 to 100 km/h. The first section was used to obtain the mathematical models for vehicle calibration and the second section was used to validate the models. Relating the RMSVA and IRI data obtained for the first section, a strong correlation between these variables was verified for the operational speeds of 40, 60 and 80 km/h for both vehicles used. From the strong correlation between these variables, 24 simple and multiple regression models were formulated and tested for each type of vehicle. Among the criteria established in the analysis, only one model was selected for the ix35 SUV and two models for the HB 20 vehicle, with simple regression models for both vehicles. Next, the reference IRI of the second stretch was compared using the Level and Rod method with IRI estimated by the selected models, considering the respective lower and upper prediction limits for each model, with 95% confidence. For the SUV ix35 vehicle, it was found that the relationship between the mean of the estimated IRI values and the mean of the reference IRI was 94%. For the HB 20 vehicle, this ratio reached, for both models, an average ratio of 97%. Therefore, it was concluded that the procedures described in this work can be used in the calibration of systems that use vertical acceleration data coming from smartphones to predict the longitudinal unevenness of the pavement.