Autonomous condition monitoring-based pavement management system

Abstract

Due to high operation cost of dedicated inspection vehicles, conventional pavement management systems (PMS) suffer from limited data quantity collected from periodic inspections. However, increasing market penetration of connected autonomous vehicles (CAVs) offers opportunities to monitor pavement conditions more frequently through sensors, including vision cameras and accelerometers, originally installed for autonomous driving. In this paper, we proposed an autonomous condition monitoring-based pavement management system (ACM-PMS) with real-time data collection using CAVs traveling voluntarily. We presented a novel mathematical framework to evaluate potential benefits of ACM-PMS in reducing social costs for both users and agency, systematically accounting for its unique three advantages: (i) large amount of condition data increases prediction model ac-curacy; (ii) aggregated measurement of current facility condition improves inspection accuracy; (iii) agency can perform maintenance activities at optimal timings, achieving continuous-time and condition-based policies. Results of numerical examples confirm that ACM-PMS significantly reduces the social cost of conventional PMS.