In many urban and for that matter rural areas, there are requirements to provide highway traffic monitoring to facilitate effective congestion management, incident detection, and highway use data collection. Even without incidents to cause significant perturbations, traffic flow is highly time varying, nonlinear, and transient in nature. Sampling the highway system at one spatial point (sensor station) at a given time generally will not provide timely situation information relative to another spatial point in the system unless the points are very close together . For systems as complex and dynamic as highways, an approach of sparse spatial sampling and use of elaborate models to provide (predict) timely situation information does not appear to be appropriate. For accurate and timely highway situation information and incident detection to be possible, traffic monitoring sensors must be spaced reasonably close together. Depending on the area (urban or rural) and the highway, desired sensor station spacing varies from 1/4 mile to several miles. Sensor station spacing of 1/3 to 1/2 mile seems to be common in many urban areas.

Deploying a high density of traffic monitoring sensor stations presents both financial and logistical problems for DOTs and municipalities. The following areas require significant consideration:

1) Hardware Procurement Costs and Life Cycle Costs,
2) Installation and Maintenance Labor Costs and Scheduling,
3) Lane Closure Impact to the Motoring Public,
4) Availability of Site Services (power and communication),
5) Intrusion into Existing Services (power, communications, and structures),
6) End User Operational Considerations (TOC data interfaces, processing, and display capability).

A multiple sensor station example is considered where sensors are deployed at 1/4 mile to 1/2 mile spacing in both directions on a multi-lane interstate highway (below). Per lane traffic flow data (such as vehicle volume, lane occupancy, and average speed) is required at real time reporting intervals (20 sec to 2 min are common) to meet traffic management requirements.

The deployment architecture centers around a single, centrally located Data/Information Concentrator (SmarTek Systems’ T-BOX) which gathers traffic flow data from each remote sensor station. The T-BOX controls the multi-point wireless network by functioning as a base station which polls each remote for its traffic data packets according to a specified reporting interval (i.e. 20 sec, 30 sec, 1 min, etc.). The geometry in Figure 1 shows 10 remote sensor stations. The number of remote sensor stations is only limited by the wireless link ranges and the desired reporting period (short reporting periods ( 1 or 2 seconds) will limit the number of sensors which can be reliably polled due to communication link overhead). We have demonstrated reliable operation in adverse RF environments for link ranges up to 1 mile.  Depending on the RF environment and geometry greater link ranges can be supported.

The T-BOX as the central Data/Information Concentrator also serves as the interface point (POTS modem, fiber, copper, etc.) to the Traffic Management System and thus eliminates redundant cost (cabinets, controllers, concrete, interfaces, labor, etc.) and logistical problems (install and maintenance scheduling and cost) associated with having a system interface (cabinet) for each traffic monitoring station. The cost savings is quantified later in this report. While not specifically discussed in this report, the use of T-BOX as the Data/Information Concentrator provides built-in capability to perform significant local incident detection processing for automatic alert generation and reporting coupled with integrated color image capture capability for up to eight cameras.

The SAS-1 Autonomous Traffic Monitoring Station (SAS-1 ATMS) (Figure 2) is configured to address all of the deployment issues listed above. The enabling technologies embodied by the SAS-1 ATMS include the following:

1) a multi-lane, non-contact, passive acoustic sensor (no pavement or motorist intrusion) installed on existing structures over the shoulder (no lane closures) or over the lanes,
2) Small lightweight components which are installed using Band-It (no structure intrusion),
3) Very low power consumption to facilitate battery power with a solar panel for charging, thus eliminating system intrusion (no need to tap into electrical service),
4) Wireless Spread Spectrum (2.4 GHz) communication (affordable and reliable) to eliminate cost, labor, and schedule issues associated with the installation of home run cables and conduit,
5) Built in sensor addressing, processing and communication capability to facilitate the use of a single centrally located data concentrator and system interface for a large number of sensors,
6) System Integration know-how and multi-point communication software resulting in easy, rapid installation (20 minutes to 1 hour per site depending on conditions).

The SAS-1 ATMS includes four basic subsystems which arrive at each site ready to install:

1) a SAS-1 (with a 2.4 GHz Spread Spectrum transceiver and antenna surge protector installed inside) for multilane traffic monitoring mounted on an existing structure over the shoulder (no lane closure) or over lanes if desired (3 to 5 lanes of coverage depending on mounting geometry),

2) a whip or yagi antenna to be positioned on the existing structure for optimum performance,

3) a small NEMA strap on enclosure which houses the solar charger and battery,

Comparison of Procurement Cost for SAS-1 ATMS Traditional Loop Detector Stations

For the multiple sensor station example being considered, each loop detector station would consist of the following components or subsystems:

1) Loop detector cabinet, controller, card file, etc.               $5400.00 to $5900.00
2) Loop detectors for three lanes                                          $350.00 to $500.00
3) Site preparation -concrete pad, pull boxes, conduit, etc.  $3000.00 to $5000.00
4) System interface - POTS modem, fiber modem, etc.          $300.00 to $500.00
  Total Estimated Cost Per Site (low to high )                       $9050.00 to $11900.00

For N traffic monitoring sites the total estimated procurement cost is N*$9050.00 (low side) and N*$11900.00 (high side). Figure 3 shows the comparison of estimated procurement cost as a function of the number of sites.

Note that there is approximately 30% difference between the estimate on the low side and the estimate on the high side. Also note that the above estimate does not include a significant labor component involved in loop detector installation support such as traffic management labor for lane closures, labor supporting pavement cuts, loop wire install into the cuts, and sealing of the pavement cuts. Considering these variations in cost and significant variation in quality and capability of loop detector installers, points to an important consideration during system procurement. Installation of loop detectors or other "in-pavement" detector technologies are typically dominated by significant and highly variable labor components. This component also suffers from continuing upward price pressures (labor rates don’t typically go down). Therefore, even if hardware costs go down, the dominant component, labor will probably go up. While the above estimates are used for comparison, evidence and experience has shown that actual per site costs could be much higher than the ones used here.

Another factor considered paramount when installing traffic monitoring stations is the installation time relative to lane closures. In most urban areas, there is no good time to close a lane. Almost always, closing lanes lead to severe congestion which leads to significant safety concerns. For any traffic monitoring technology, the goal should be to not only minimize or eliminate the necessity of lane closures but to minimize the total install time required at each site. Experience has shown that typical install times for loops on major highways is measured in terms of at least several hours for each site.

For the multiple sensor station example being considered, each SAS-1 ATMS would consist of the following components or subsystems:

1) SAS-1 for multi-lane monitoring (3 to 5 lanes)                                                           $3500.00
2) Solar Power subsystem (30 watts) including panel, NEMA enclosure, battery, charger $850.00
3) Spread Spectrum wireless subsystem including transceiver, surge protector, antenna $800.00
                                                                            Total Estimated Cost Per Site       $5150.00

For the multiple sensor station example being considered, the centrally located Data/Information Concentrator would consist of the following components or subsystems:

1) Small strap-on NEMA enclosure, power converter, and battery $450.00
2) T-BOX processor                                                                $2500.00
3) Spread Spectrum wireless subsystem including
    transceiver, surge protector, and antenna                                $800.00
4) System interface - POTS modem, fiber modem, etc.                $500.00
                                                          Total Estimated Cost     $4250.00

For N traffic monitoring sites the total estimated procurement cost includes one central Data/Information Concentrator subsystem ($4250.00) + N-SAS-1 ATMS (N*$5150.00). Figure 3 shows the comparison of estimated procurement cost as a function of the number of sites. Note that the cost estimates above are approximate and for comparison purposes only. These estimates do not represent a quote.

The SAS-1 ATMS components will arrive at the each site ready to be installed on existing structures using Band-It. Therefore, install time is minimized and correspondingly install labor is minimized. Required lane closures are eliminated. Typical install times should be no more than 1 hour and could be as low as 30 minutes per site. The install labor component for SAS-1 ATMS is by design a small contributor to station cost. As compared to loop detector sites, the dominant component for SAS-1 ATMS sites is the procured hardware. This hardware is continually under downward price pressure. It is therefore, reasonable to expect that future costs for SAS-1 ATMS to hold steady or go down. Quantity buying can immediately realize reduced costs.

Summary