Roadside Truck Placard Readers for Advanced Notice and Response at Safety-Critical Facilities: Phase 2

About the report:

The transport of hazardous and dangerous materials (HAZMAT) through safety-critical facilities poses significant risk to overall system reliability should those assets be incapacitated by the occurrence of a related incident. This problem is particularly acute for facilities in remote locations such as Virginia’s mountain tunnels on Interstate 77 (I-77) due to limitations on alternate routes and the availability and proximity of emergency responders and specialized equipment/supplies. Automated placard reader systems (APRSs) are commercially available camera-based computer vision systems that “read” hazardous material placards on passing trucks from roadside installations. This information, along with other pertinent vehicle identification data may then be forwarded to critical facility operators to inform any preparations or responses that may be required.

The Virginia Tech Transportation Institute conducted an initial phase of work to assess the readiness of APRSs for their reliable and effective roadside deployment and to determine how the data from such a system could be used by facility operators to improve safety and mitigate disruption during an event involving HAZMAT. The findings of the first phase of work indicated that available APRS technology was sufficiently advanced to warrant a second phase of work that included field testing and further refinement of the preliminary deployment plan. In Phase I, an APRS from Intelligent Imaging Systems (IIS) was identified for further evaluation. In this (second) phase of work, a mobile APRS system provided by IIS was evaluated under experimental and naturalistic scenarios at the Virginia Smart Roads and at several locations on Virginia public roads. A photographic survey of public HAZMAT placard usage conducted previously was used to inform this testing. Additional naturalistic data were acquired from a permanent APRS installation in Delaware when difficulties with the mobile APRS were encountered.

The mobile and permanent APRSs were able to classify HAZMAT placard accurately at rates of 96% and 99%, respectively. The mobile and permanent systems were able to read United States Department of Transportation (USDOT) numbers from the sides of tractors correctly at rates of 46% and 67%, respectively, and tractor license plates correctly at rates of 43% and 39%, respectively. Moderate levels of rain and snow, as observed through roadside cameras and reported at nearby weather stations, had minimal impact on reporting accuracy. Performance of the system at night compared favorably with daytime performance. To address potential false negative concerns, a visual survey of 187 commercial vehicles was conducted that revealed that the APRS was 85% successful at locating and identifying the presence of placards on commercial vehicles passing in the near lane. With respect to implementation, the nature of how the subject APRS data is provided to users is not currently conducive to automated integration with existing or future VDOT tunnel or traffic management systems as data must be read from an online interface and no “push” options are currently available. Also, providing advance warning of the approach of HAZMAT to tunnel operators on I-77 is not feasible given constraints related to the geographic siting of potential APRS installations and respective traffic characteristics. However, facility operator access to APRS data after an incident has occurred may provide benefits of improved responder and traveler safety as well as faster clearance times.  

Authors