SIMPL-Tolling
110-Day Independent Field Validation
In a 110-day live deployment at a North American all-electronic tolling (AET) facility, SIMPL analyzed more than 1.38 million transactions.
It reached 100% accuracy in vehicle classification, length and height, and cut the reconciliation ghost rate to 0.00%.
At a glance
The challenge
Toll operators depend on their roadside and in grantry detection technology to capture and classify every vehicle. The same detection also acts as the rigger point to activate their cameras for front and back license plate recognition.When a transaction goes missing or a vehicle is misclassified, the revenue loss is silent. It only shows up later, if at all.
Operators need more than a claim of accuracy. They need an independent record of every vehicle, reconciled and integrated against their BOS, so exceptions surface before they become revenue issues.
The deployment
SIMPL was installed at a live North American AET facility and ran alongside the existing tolling system for 110 days. SIMPL independently detects, measures every vehicle crossing and creates a transaction-level record that can be reconciled against existing tolling systems.
- Detect: Seyond LiDAR sensors captured 3D data of every vehicle crossing the tolling zone.
- Measure: SIMPL's AI perception software recorded each vehicle's FHWA 13-class category, length, height, speed, lane position and timestamp.
- Record: Each crossing became a transaction-level record with a unique identifier.
- Reconcile: The concessions owner matched SIMPL records against their live BOS to flag missing, extra or misclassified transactions.
- Consistency: Know every vehicle is detected and tiggers cameras according to speed, heading, trajectory and lane
The 110 days included 26 rain days, testing performance in the weather operators actually face.
Results: measurement accuracy
By the end of the evaluation, SIMPL measured vehicle classification, length and height with 100% accuracy against independent ground truth. Accuracy improved as the models were tuned on site, then held at that level for the final six weeks of operation.
For operators, this means the classification behind each toll rate can be checked against a source they can trust.
Results: reconciliation performance
SIMPL was also tested as a reconciliation platform against the operator's live BOS. Both exception rates fell sharply over the evaluation, and ghost transactions dropped to zero.
The remaining orphan transactions were traced to documented road works and operational conditions, not sensor or classification failures. In other words, SIMPL was flagging real-world events the operator would want to know about.
What it means for operators
Accuracy alone isn't enough. The value comes from checking every tolling transaction against an independent source of truth and catching exceptions before they become revenue issues.
- Revenue Assurance: Find missed transactions and classification errors before they hit revenue.
- Independent Audit Layer: Works alongside the existing BOS through standard interfaces, with no replacement needed.
- Less Manual Effort: Daily exception reports focus staff on the transactions that need attention.
- More Accurate Zone Control: Trigger your cameras to capture front and back license plates, even on vehicles going up to 200km/h (135mph)
Interested in learning more, request more information here.