Application of an Inexpensive Sensor in Calculating the International Roughness Index

Abstract

The international roughness index (IRI) is one of the most common indices applied in the assessment of road roughness. The initial step in calculating the IRI is to collect depth data from a road surface. Depth data is commonly collected using an automated data collection vehicle. This vehicle has some advantages, such as a higher level of safety, precision, accuracy, and repeatability as compared with the manual data collection method. However, conventional automated data collection methods are of significant cost to purchase, operate, and maintain. A trade-off between quality and cost in roughness data collection has not been studied sufficiently. The main purpose of this study is to propose a sensor that is inexpensive and of sufficient quality to collect data for IRI calculation. The proposed sensor, Kinect V2, has the capability of collecting color and depth images from a road surface. Using these images, a three-dimensional (3D) model of the road is built. The IRI is finally computed through the application of this model. The results are successfully validated through application of an accurate manual device.