The United States is suffering from an aging civil infrastructure crisis. Key to recovery are rapid
inspection technologies like that being investigated by the VOTERS project (Versatile Onboard
Traffic Embedded Roaming Sensors), which aims to outfit ordinary road vehicles with compact
low-cost hardware that enables them to rapidly assess and report the condition of roadways and
bridge decks free of driver interaction.
A key piece of hardware, and the focus of this paper, is a 24 GHz millimeter-wave radar system
that measures the reflectivity of pavement surfaces. To account for the variability of real-world
driving, such as changes in height, angle, speed, and temperature, a sensor fusion approach is
used that corrects MWR measurements based on data from four additional sensors. The corrected
MWR measurements are expected to be useful for various characterization applications,
including: material type; deterioration such as cracks and potholes; and surface coverage
conditions such as dry, wet, oil, water, and ice.
Success at each of these applications is an important step towards achieving the VOTERS
objective, however, this paper focuses on surface coverage, as whatever covers the driving
surface will be most apparent to the MWR sensor and if not accounted for could significantly
limit the accuracy of other applications. Contributions of the paper include findings from static
lab tests, which validate the approach and show the effects of height and angle. Further
contributions come from lab and in-field dynamic tests, which show the effects of speed and
demonstrate that the MWR approach is accurate under city driving conditions.