Route Points Analyzed--
Across route geometry
Real route grounding
Traffic systems count vehicles. PTIS predicts where they are going.
A vehicle that keeps crossing checkpoints without turning is giving destination evidence. PTIS converts those movement patterns into early destination belief, helping predict bottlenecks before they fully form.
Software validated
Real route grounded
Pilot-ready
Route-Grounded Prediction
Silk Board→Whitefield
PTIS belief after checkpoint
--Whitefield
Loading latest replay decision
Observation Events--
Processed in replay
Prediction Confidence--
Toward Whitefield
Safety Violations--
Capacity gate result
How PTIS works
Click a junction and watch the belief change.
Each checkpoint crossing without a turn removes one destination possibility. PTIS treats that absence as evidence, then applies a capacity gate before any software action becomes eligible.
Confidence: vehicle is heading to Whitefield
18%
65% threshold: connector is only eligible after the capacity gate is safe.
Vehicle enters at Silk Board
PTIS starts from a prior belief. Six possible destinations remain, so the system observes without acting.
Junction 1 / 6
The idea in one flow
Do not wait for the jam to become obvious.
1
Observe
Use checkpoint counts or anonymized linkage tokens. No GPS path, no number plates, no face data.
2
Eliminate
Every passed junction removes one possible destination from the belief set.
3
Gate
Activate only when belief is high and the connector has safe capacity.
What is proven today
Software proof, not field-deployment claim.
Scenario checks passed.
Vehicles replayed through the decision logic.
Observation events processed.
Capacity gate violations found.
Images / boxes audited.
Capacity reference basis.
Truth boundary
What this page is allowed to claim
Claimed
Synthetic OD calibration is loading. Replay software works on deterministic scenarios, stress load, real route geometry, public CCTV detection evidence, and official planning references.
Not claimed yet
No live BTP deployment, no Google Maps integration, no field-measured congestion reduction, no real destination-ground-truth validation, and no proven field re-identification method yet. Checkpoint gaps and slip roads remain pilot-design risks.