Smartphone Technology Integrated with Machine Learning for Airport Pavement Condition Assessment

Abstract

Safe and efficient aircraft operation requires airfield pavements to possess adequate load carrying capacity, smoothness, and should be free of foreign object debris (FOD). To remain free of FOD, the type, extent and severity of several key distresses (e.g., various cracking forms) must be monitored and, as necessary, maintained and repaired. Although obtaining data regarding airport pavement condition is an essential task for pavement maintenance, and for prioritization of rehabilitation and reconstruction activities, pavement condition surveys can be costly and labor intensive. This study presents a potential low-cost and efficient approach to assess airport pavement condition based on the use of modern smartphones. A smartphone application was deployed to record vehicle cab acceleration data, timestamp, and GPS coordinates. The approximate airport pavement profile was then back-estimated from the acceleration data using an inverse state-space model. The model considers the physics of mass-spring-damper system of the vehicle sprung mass. The analyses were performed using a MATLAB script to calculate pavement international roughness index (IRI) values. The initial phase of this study calibrating the system on a designated test section along MO-10E near Excelsior Springs, Missouri in order to locally calibrate model parameters. The smartphone-measured IRI values were then compared with known IRI values measured by an Automatic Road Analyzer (ARAN) van as a means of validation. Subsequently, the validated smartphone application was used to collect acceleration data for airport pavements at the 26 state funded general aviation airports in Missouri. Each test run was conducted multiple times along the centerline and in presumed gear offset locations along the length of the 26 general aviation airport (GAA) runways studied. The smartphone-based IRI values for the airports were found to be in close agreement with the construction and maintenance records obtained for the runways investigated. A pavement condition index (PCI) rating system is widely used in the US for the assessment of the overall condition of airport pavements. A robust machine learning technique was developed and shown to be a reasonable estimator of foot-on-ground reference PCI ratings, using only smartphone-measured IRI as an input. The work is being extended to include basic condition rating information, which can be easily collected by airport managers and maintenance personnel.