Tennessee now has four sensor methods watching the same parking problem from different angles. At the Interstate 40 Welcome Center in Smith County, the Department of Transportation has made a truck-parking availability pilot operational and connected it to SmartWay. Magnetometers, radar-camera fusion, LiDAR and closed-circuit cameras are all part of the test.
TDOT says it will evaluate the methods over the coming months, then recommend technologies and standards for possible statewide scaling across truck-parking facilities. That makes this more than a traffic-information trial. It is a public test of how Tennessee chooses sensing equipment, verifies the resulting data and limits what cameras may do. Read TDOT’s September 2 project release.
For security directors, parking operators and companies selling occupancy technology, the immediate lesson is practical: a live count on a sign is the end of a chain, not proof that the chain works. Before any wider purchase, the state needs evidence about accuracy, stale data, outages, access and maintenance. Private buyers should demand the same evidence.
This is a comparison, not a statewide rollout
The Smith County equipment is installed and operating. Statewide deployment is still a possible future step.
That distinction matters because the September 2 release does not announce a selected sensor, a statewide contract, an expansion budget or an accuracy result. It says QR codes at the parking area invite drivers to comment on the technologies. TDOT will use the evaluation to determine recommended technology and system standards for future scaling.
The federal funding record also describes a test stage. The U.S. Department of Transportation lists a $1,986,000 award to TDOT for the Tennessee Truck Parking Availability System and classifies the project as a sensor project addressing truck parking in Smith County. That figure is the federal award for the project. It is not a published unit price, the cost of any one detection method or an approved statewide budget. Check the FY2024 SMART Stage 1 project list.
The structure of the SMART program explains why a result can include a decision to stop. The federal program’s fact sheet says Stage 1 recipients should finish with performance evidence for an implementation report or enough information to decide against proceeding. It says a prototype may expose technical limits, institutional barriers or weak performance compared with conventional options. See the U.S. DOT SMART program fact sheet.
A good pilot does not owe anyone a rollout. It owes decision-makers a defensible result.
The four methods are not described as four identical tests
TDOT’s short release reveals a detail that future performance reporting will need to handle carefully. The four technologies do not all appear in the same measurement arrangement.
Magnetometers count trucks entering and leaving the eastbound lot. The radar-camera fusion sensor also counts eastbound arrivals and departures, while also detecting the number of available spaces in the westbound lot. LiDAR and CCTV detect the number of available spaces in the westbound lot.
That setup can answer several questions. It can compare two ways to estimate occupancy from vehicle flow on the eastbound side. On the westbound side, it can compare three methods for producing an available-space count. What the public release does not provide is the scoring plan, test period, ground-truth method or acceptable error rate.
That silence is not evidence that the project lacks those controls. A news release is not an evaluation plan. It does mean readers cannot yet rank the technologies or declare a winner.
A publishable evaluation should make the comparison legible. How was the actual number of open spaces established? Were errors counted by hour, day and operating condition? How quickly did each method correct after a missed entry or exit? Did a false “space available” result carry more weight than a false “full” result because a driver might continue toward a lot with no place to stop?
Those are editorial questions, not claims about unpublished test results. They are also the questions a warehouse, distribution center, private truck stop or guarded campus should put into an occupancy-system acceptance plan. Vendor demonstrations tend to show a clean view and a cooperative vehicle. Acceptance testing should use the site that will actually be monitored.
TDOT drew a clear camera boundary
The release includes one direct privacy rule: any camera used at the Smith County site to monitor parking will not be recorded or used for enforcement. TDOT presents that treatment as similar to its SmartWay cameras.
That sentence is useful because it defines two prohibited uses in public. A future specification can turn the statement into testable requirements: recording stays disabled, and parking-monitoring imagery does not become an enforcement feed.
The release does not describe every other part of data handling. It does not state what occupancy records or equipment logs are retained, who can administer the sensors, whether a contractor can view a live feed, how access is logged, or how a later change in purpose would be approved. None of those omissions proves that a safeguard is absent. They mark the boundary of what the September 2 public account establishes.
Before statewide expansion, TDOT should publish or point to the rules that govern the rest of the chain. At minimum, a buyer needs to know what data leaves each device, where it goes, who can retrieve it, when it is deleted and how unauthorized access is investigated. If images are processed without being recorded, the system description should say whether they ever leave the device and what derived data remains.
Private security buyers often receive a broad promise that a system “doesn’t store video.” That answer is too narrow. An occupancy platform may still create timestamps, device identifiers, alerts, health records, administrator logs and exports. A contract should name each data type and its retention rule instead of treating “video” as a substitute for a data inventory.
Real-time guidance needs a failure state
The pilot sends availability information to drivers through roadside signs and in-cab alerts on electronic logging devices. The state says those notices can help a driver decide whether to continue toward the Welcome Center or seek parking earlier.
Once a driver relies on that count, data quality becomes an operational safety issue. The system needs a defined answer for a sensor that stops reporting, a connection that fails or a count that has not refreshed. “Unknown” is a legitimate state. Presenting an old number as current is not.
This is where a parking pilot becomes relevant to security operations. Guard tours, gate status, visitor counts and perimeter alarms also pass through sensors, networks, software and displays. A dashboard can look healthy after the underlying device has gone quiet unless the design carries device health and data age all the way to the operator.
An acceptance test should therefore include ordinary failures. Disconnect a sensor. Interrupt its network path. Restore power. Check what the sign, in-cab alert and any operator display show during each step. Record the time required to detect the problem, suppress an unreliable count and return to service without duplicating or losing events.
The test should also separate availability from integrity. A service can remain online while showing a wrong count. Teams need a manual way to verify the lot, correct the displayed value and record why an override occurred. Administrator changes should leave an audit record. Credentials should belong to named roles, and a departing employee or vendor technician should lose access on a defined schedule.
TDOT’s separate Intelligent Transportation Systems Qualified Products List page supplies a useful caution, though it does not say the QPL governs this pilot. The page says listed products still must complete required testing, integration, approval and acceptance. TDOT may reject or remove a product that performs poorly under real-world conditions. Review TDOT’s ITS product process.
A catalog entry is not an acceptance result. Neither is a successful installation.
Scale the standard only after the evidence
TDOT has already made one sound choice: it is comparing methods before naming a statewide answer. The next test is whether the public result preserves enough detail for another Tennessee site to understand the choice.
The evaluation should report the observation period, actual-versus-reported counts, latency, outages, maintenance visits and driver feedback for each configuration. It should distinguish eastbound flow counting from westbound available-space detection. Costs should separate equipment, network service, software, calibration, repair and replacement. The privacy record should identify data retained by the state and by each contractor.
A statewide standard also needs room for different sites. A method that works at one welcome center may face a different view, pavement, traffic pattern, power source or communications link elsewhere. The answer may be a defined menu with site-selection criteria rather than one device for every lot. That is an issue for the evaluation to settle, not a conclusion supported by the release today.
Security and parking managers do not need to wait for TDOT’s final report to improve their own buying process. They can require a pilot with written success thresholds, observe it against a manual count, test its failure state, inventory its data and make final payment depend on acceptance evidence. They should also preserve the right to remove or replace equipment that fails after installation.
Tennessee’s pilot now tells drivers whether parking is available at this Welcome Center. The pending evaluation still has to show how well that answer performs and whether the technology and standards should scale.
TN Security Review is an independent publication. It is not affiliated with TDOT, the U.S. Department of Transportation, SmartWay, or any vendor discussed in this article.