A truck parking availability information consisting of dynamic message signs (DMSs), a mobile application, a public Web site, and an interactive voice response system (IVR) to disseminate the real-time information to truck drivers is tested Tennessee.
A public rest area near Athens, TN located at mile marker (MM) 45 on I-75 Northbound (N) was selecting for the testing. An additional site was added at a non-operational weigh station at MM23 on I-75 N, 22 miles from the first site.
The design of the system was comprised of several significant system components, described below:
- Detectors: Side-mounted infrared laser scanners used to detect vehicles at ingress and egress.
- Doppler radars: Used to provide velocity and length information of a vehicle relative to the laser scanner. Vehicle length was used to classify vehicles
- On-site processor: Used to process the scanner and light curtain signals.
- Off-site server: Used to download and store data in a database.
- Closed-circuit television (CCTV) cameras: Used for validating site space availability (ground truth).
- Web site and data archive: Used to monitor the CCTV cameras and make corrections to the site as needed.
- Dynamic message signs (DMSs): Utilized to disseminate parking availability information to users on the roadway.
- Interactive voice response (IVR) system: Truck drivers were able to call 1-844-SMART-PK to check parking availability at the site and reserve a parking space.
- Mobile application: Truck drivers were able to use the application on a smartphone to check parking availability at the site and reserve a parking space.
Some of the challenges, constraints and drawbacks of the project’s approach are described in the lessons below.
Detector accuracy is heavily dependent upon proper installation. The detectors, if installed properly, are able to accurately identify vehicles entering and exiting a site. However, due to variations in the installation of the detector technology between the ingress and egress of a site, a vehicle leaving the lot may have been classified differently than it was classified when entering the lot. The "check-in-check-out" approach to detecting parking occupancy resulted in any error from the detectors compounding over time. Even small error rates can result in the traveler information component disseminating incorrect information after 1–2 days without intervention. For this reason, the system must be calibrated at least once per day to maintain acceptable accuracy levels.
The detectors are also highly sensitive to orientation with respect to the roadway. Manufacturer recommendations should be followed precisely to ensure effective detections. During the testing period the orientation of one detector was altered due to impact by a truck. After this impact the detector accuracy was reduced and the system needed to be taken offline until the orientation could be rectified.
Fixed closed circuit television (CCTV) cameras were utilized for the surveillance of the parking area and ingress and egress points and to assist in evaluating the performance of the sensors. The cameras presented several issues during operation regarding line of sight. Future deployments will ideally replace all the static cameras with pan/tilt/zoom (PTZ) cameras, also reducing the number of cameras needed. Improved cameras would make calibrations much easier by allowing operators to view more of a lot, zoom in and out to examine areas of occlusion, and obtain a much more comprehensive view of all areas of a lot to identify any abnormal activity.
Installing a single, high-mounted, PTZ camera would also eliminate the need for the numerous cameras used to perform calibrations in the field operations test. This would reduce ongoing bandwidth costs to the site and reduce maintenance costs by minimizing the quantity and type of field equipment.
Vehicle Classification Scheme
The system’s detectors were programmed to classify vehicles based on length and place them into bins. Early in the project, results showed the limitation of classifying vehicles based on length, as the accuracy rate across the original six bins of vehicle lengths was low. Length was used again later in the project but with a reduced number of only 4 bins, each covering a different range of vehicle lengths ("small vehicles", "large vehicles", "oversized vehicles", and "bobtails". Reducing the number of bins made the classes more easily discernible from one another to anyone in the field attempting to conduct counts to corroborate or evaluate the system.
The system’s classification ability could also be a valuable source of baseline information regarding the site. The detectors, for example, can somewhat inform an operator when a bobtail enters the lot. This is useful to know when the site is being utilized for trailer drops. The system could also inform operators when an oversized load enters the facility, allowing them to check logs for permits and to potentially issue safety warnings to the vehicle if conditions have changed since they obtained the permit. The vehicle classification scheme would have very limited value in trying to fit two bobtails into a regular truck parking space in order to maximize using the parking space capacity of a site.
A reservation system was implemented at both parking facilities. An implicit honor system was in effect for a reservation—if a trucker did not have a reservation, he was expected to park in an unreserved space. Reservations were not patrolled or managed in any way. With no incentive to observe the rules of the reservation system, truckers simply did not use it. The reservation component should be removed from further SmartPark deployments unless the system owner commits to all aspects of the service, including monitoring of the facilities and enforcement of the rules.
Traveler Information Utility
The traveler information system built for the deployed system included DMSs, an IVR system, a mobile application, and a public Web site. Data collected on usage of the various components showed that the most useful aspect of the traveler information system was by far the DMS, which reached all passing motorists and was strategically placed to maximize the utility of the information. To use the mobile application and to make reservations or check historical data on the Web site, users had to register. Participating truckers generally did not download the mobile application, and hits on the public Web site were limited.
Although it was clear that the DMSs reached a much wider audience, mobile applications and Web sites should not be disregarded or marginalized as traveler information tools. The relatively low usage of the SmartPark e-kiosk components can likely be attributed to several factors, including the lack of general availability of real-time truck parking information. Truckers were not used to the availability of such data and as a result were less likely to adopt and trust it. The SmartPark traveler information system components were also not broadly advertised (advertising was not within the scope of the project). A strong marketing campaign could potentially have boosted awareness and use of the system.