State departments of transportation are utilizing unmanned aircraft systems (UAS) in a variety of applications that support planning, construction, maintenance, and operations within their jurisdictions. As the use of UAS to support road-traffic-related applications is currently limited by a host of technical, regulatory, and operational issues, further exploration of these aspects is needed to inform the Virginia Department of Transportation’s (VDOT’s) plans to expand the use of UAS to improve travel safety and traffic operations. This exploration included conducting a comprehensive literature review to determine the current state of technology and practice, developing a conceptual plan for UAS implementation into VDOT operations, and performing a UAS application pilot demonstration. The state-of-practice review included VDOT’s use of UAS, as well as that of other state DOTs that have employed UAS in their operations. As part of the implementation plan development, the researchers conducted a survey of state transportation agencies to assess the level of UAS usage, the types of applications that have been used, and impediments to implementation.
This work indicates that 36out of 50 states (72%) are currently employing UAS for various transportation applications, and many are funding centers and programs for UAS operations. Moreover, each of the state DOTs is approaching the introduction and use of UAS technology differently. In many cases, states that had an advanced UAS program have also had a champion who achieved early adoption by leveraging available resources, knowledge, and experience. The experience of these agencies varies widely; some have a wide range of UAS application experience, while others have explored a much narrower range of uses. A preliminary implementation plan was developed that encompassed aspects of potential UAS traffic applications and metrics that could be used to assess an effective UAS deployment.
After review of numerous potential UAS applications, assessment of traffic flow at intersections was selected for pilot demonstration. UAS video data was collected at five intersections, and the data from one intersection was analyzed using an online computer vision tool to characterize traffic flow. This intersection was chosen for analysis due to the availability of data from an existing camera-based traffic monitoring system. The two methods of traffic video analysis were described and compared using a small segment of live traffic data. The existing traffic monitoring system uses cameras mounted on signal mast arms and relies upon highly oblique views of vehicles across multiple lanes of traffic, which often results in visual occlusion. In general, the improved view afforded by UAS data acquisition translated to a more accurate assessment of traffic flow. UAS acquisition also provides more flexibility with respect to deployment and a better, more expansive view of the intersection and approaching traffic. However, the UAS system is very limited with respect to viewing time as battery-powered flights are short, and flights over people and moving vehicles are currently restricted by regulations. Longer flight data collection may be accomplished using tethered UAS where power is provided to the aircraft directly from the ground.