Traffic Flow Benchmarking and Intersection Control

AI traffic management suite for congestion reduction, combining multi-scale traffic forecasting, realistic gap-aware benchmarking, and cooperative intersection trajectory prediction to improve planning, evaluation, and safer flow control.

The Problem

“Reduce congestion and improve safety with resilient, multi-scale traffic forecasting and cooperative intersection prediction”

Organizations face these key challenges:

1

Traffic forecasts often perform differently at hourly versus daily aggregation, making model selection difficult

2

Evaluation datasets are unrealistically clean compared with production sensor environments

3

Sensor failures, maintenance swaps, and communication outages create missing-context intervals

4

Single-view intersection prediction misses intent and interaction cues from other agents and infrastructure

Impact When Solved

Improve forecast usefulness for both hourly operations and daily planningBenchmark traffic AI under realistic sensor outages and missing-context intervalsIncrease robustness of deployed models when data feeds are incomplete or delayedEnhance intersection safety with cooperative multi-view trajectory prediction

The Shift

Before AI~85% Manual

Human Does

  • •Review historical traffic reports and choose planning assumptions for hourly and daily operations
  • •Compare forecasting results manually across time scales and decide which model outputs to trust
  • •Investigate sensor outages, maintenance changes, and missing data before using results
  • •Assess intersection behavior from single-view evidence and decide on safety or control actions

Automation

  • •Produce basic traffic forecasts from historical sensor data
  • •Generate standard performance reports on mostly clean continuous datasets
  • •Flag obvious data gaps or feed interruptions in incoming traffic records
With AI~75% Automated

Human Does

  • •Approve which forecast horizon and aggregation level should guide planning or operations
  • •Review benchmark results under outage conditions and decide model deployment readiness
  • •Authorize signal, routing, or safety interventions recommended by the system

AI Handles

  • •Generate hourly and daily traffic forecasts and compare expected usefulness for different decisions
  • •Create realistic missing-context benchmark scenarios and score model robustness under degraded data
  • •Monitor live feeds for outages, delays, and context loss and adjust confidence in outputs
  • •Predict cooperative intersection trajectories from infrastructure, vehicle, and shared views

Operating Intelligence

How it works

AI runs the first three steps autonomously.

Humans own every decision.

The system gets smarter each cycle.

Confidence91%
ArchetypeRecommend & Decide
Shape6-step converge
Human gates1
Autonomy
67%AI controls 4 of 6 steps

Who is in control at each step

Each column marks the operating owner for that step. AI-led actions sit above the divider, human decisions and feedback loops sit below it.

Loop shapeconverge

Step 1

Assemble Context

Step 2

Analyze

Step 3

Recommend

Step 4

Human Decision

Step 5

Execute

Step 6

Feedback

AI lead

Autonomous execution

1AI
2AI
3AI
5AI
gate

Human lead

Approval, override, feedback

4Human
6↺ Loop
AI-led step
Human-controlled step
Feedback loop
TL;DR

AI handles assembly, analysis, and execution. The human gate sits at the decision point. Every cycle refines future recommendations.

The Loop

6 steps

1 operating angles mapped

Operational Depth

Technologies

Technologies commonly used in Traffic Flow Benchmarking and Intersection Control implementations:

AutoformerOther
1 mentions
Benchmark scriptsOther
1 mentions
Daily aggregated traffic datasetsOther
1 mentions
DLinearOther
1 mentions
Gap data CSVsOther
1 mentions
+10 more technologies(sign up to see all)

Key Players

Companies actively working on Traffic Flow Benchmarking and Intersection Control solutions:

Argoverse APIAutoformercontinuous-history traffic forecasting benchmarksDLinearHiVT
+6 more companies(sign up to see all)

Real-World Use Cases

Multi-scale traffic forecasting with hourly and daily aggregation

Forecast traffic at different zoom levels, like hourly and daily, to understand which time scale works best for planning and analysis.

multi-scale time-series forecastingresearch / benchmarking use case based on comparative experiments rather than a production deployment pattern.
10.0

Gap-aware benchmark generation for realistic traffic AI evaluation

Create test scenarios where AI must forecast traffic after a long blind spot, like predicting road conditions after sensors were offline for a while.

forecasting with missing-context intervalsimplemented benchmark workflow with sample files and scripts; still primarily a research evaluation tool rather than a packaged production system.
10.0

Cooperative trajectory forecasting at intersections

Predict where cars and other road users will go next by watching both the road from above and the vehicle’s own view.

sequence forecastingresearch-grade but practically relevant; the repository releases dataset assets and benchmark support for forecasting workflows.
10.0

Free access to this report