Tollscopic Tolling · RTCS for all-electronic tolling

Full roadside tolling without cutting the pavement.

Tollscopic is an AI-native RTCS for all-electronic tolling: gantry-mounted sensing, trajectory-based event correlation, multi-sensor classification, immutable evidence, and clean integration with the back office already in place.

Gantry-only RTCS
Patented trajectory correlation
Auditable evidence chain
Back-office independent
01 · The legacy problem

Roadside tolling has accumulated devices faster than it has changed architecture.

Cameras, RFID readers, axle sensors, LiDAR, and lane controllers all observe the same traffic stream, but many systems still depend on event timing and device-specific chains to decide which signals belong together. That works until a real road gets messy.

Stop-and-go flow. Trucks blocking camera views. Vehicles straddling lanes. Missing reads. Curved geometry. Sensors that disagree. The result is not just a missed read. It is a transaction with weak provenance, an exception that requires manual review, or a system that cannot improve without changing the roadside again.

The architectural problem
Audit happens after the transaction has already lost its raw context.
Correlation models · side-by-side
LEGACY Timing-window matching t0 t1 PLATE TAG AXLE PLATE TAG AXLE which events belong together? – overlapping vehicles – occluded plates – missing reads TOLLSCOPIC Trajectory binding evidence bound to the path + each vehicle is one object + evidence has provenance + ambiguity becomes reviewable
02 · The Tollscopic architecture

Four layers. Evidence is the stable one.

Each roadside device produces evidence. Edge compute turns video into vehicle trajectories. The cloud transaction layer binds evidence to those trajectories, fuses classification signals, and builds the transaction record. A back-office interface publishes that record without making the roadside logic dependent on one billing platform.

EVIDENCE FLOW → 01 · ROADSIDE Gantry sensing ANPR cameras plate evidence Sidefire / wide-angle vehicle path video Axle / class sensors vision · LiDAR RFID readers transponder evidence Cabinet + WAN edge compute housing 02 · EDGE Trajectory generation – NVIDIA Jetson Orin – vehicle tracking – local buffering – device-health checks close to the road 03 · CLOUD Transaction intelligence 1 Evidence binding 2 AVDC classification fusion 3 Confidence + provenance 4 Transaction assembly 5 Adapter output AWS · event-driven · stateful 04 · DELIVER + AUDIT BOS adapter + replay – BOS / OBO interface – DVAS audit surface – immutable evidence store – replay + version trace BACK OFFICE Existing BOS / OBO stays in place
FIG · The Tollscopic tolling architecture, conceptual
01 · Roadside sensing

Gantry-mounted cameras, RFID, classification sensors, LiDAR where configured. Every device produces evidence.

02 · Edge trajectory generation

Edge compute turns sidefire and wide-angle video into a vehicle trajectory through the zone.

03 · Cloud transaction intelligence

Evidence binds to the trajectory. Classification is fused with confidence and provenance. The transaction is assembled.

04 · Back-office + audit

Transactions publish through a stable adapter to the operator's BOS / OBO. The immutable evidence store remains available for replay.

03 · Why it wins

The advantage is not a single sensor. It is how the system treats every sensor as evidence and every vehicle as a tracked object.

That gives the platform a better way to handle ambiguity, a stronger basis for audit, and a more flexible path for modernization.

Gantry-only sensing
Reduces dependence on civil works and in-road sensors
Trajectory correlation
Handles dense, ambiguous traffic better than timing windows
Multi-sensor fusion
Classification has confidence, provenance, and fallback paths
Immutable evidence
Transactions can be audited, replayed, and improved
Hardware adapters
Operators avoid single-device lock-in
BOS / OBO integration
Existing billing systems can remain in place
Reliability loop
Monitoring and replay support continuous improvement
05 · Field validation

Real roadside constraints have shaped the product.

The story below is about engineering conditions, not market footprint. Each condition is a technical lesson the system has had to absorb.

Condition
Multilane flow

Overlapping vehicles and dense evidence streams

Condition
Curved geometry

Off-axis camera views and lane ambiguity

Condition
Bridge + weather

Wind, salt, and difficult maintenance access

Condition
Cellular-backed

Degraded WAN handled with local buffering

Bring us the toll zone, the constraints, and the back office.

Tollscopic is strongest where the old architecture is becoming the bottleneck: dense traffic, brittle lane hardware, opaque audit, and systems that cannot improve without major roadside change.