Real-World Application and Validation of Vehicle-Mounted Pavement Inspection System

Pavement condition surveys are typically done manually by inspection teams. They can be expensive and can take several months to complete. This paper considers a vehicle-mounted pavement inspection system that aims to be more affordable and faster than manual surveys. The system is comprised of a vehicle outfitted with a multi-modal sensor array that includes accelerometers, microphones, radar, tire pressure, and video. In a first step, each channel is processed individually to extract parameters related to friction, roughness, and distress. The second step applies data fusion concepts to combine all sensor data into a single pavement condition value. This paper elaborates on this system and describes a real-world field test. The primary contribution is a proof-of-concept for the system by comparing field test results to those from a conventional manual survey.