Agent-Based Traffic Simulation Using Privacy-Safe Synthetic Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing traffic simulation models using synthetic data are inaccurate and lack realism, failing to replicate real-world driving behaviors, while using real-world data raises privacy concerns.

Innovation Solution

A system that uses agent-based modeling to simulate traffic by translating real-world agent-based position data into narrativized instructions, training a machine-learning model to generate synthetic movement data that mimics real-world driving behaviors, and outputs metrics associated with simulated environments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If real-world agent-based position data is used to simulate traffic, then simulation accuracy and realism are improved, but privacy concerns and data security risks worsen

Engineering Contradiction:
Improvesimulation accuracyVSAvoidprivacy concerns
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The system creates synthetic agent-based position data that copies and mimics the statistical properties and movement patterns of real-world traffic data without containing actual personal information. The synthetic data is generated through machine learning models trained on real data, preserving the essential characteristics needed for accurate traffic simulation while eliminating privacy risks associated with real individual location data.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The system introduces synthetic data as an intermediary between real-world traffic patterns and simulation models. Instead of directly using real individual position data, the system uses machine learning models as intermediaries to translate real data patterns into synthetic representations, thereby maintaining simulation accuracy while removing direct access to sensitive personal information.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Object-affected harmful factors

If synthetic data is used to simulate traffic, then privacy concerns are addressed, but simulation accuracy and realism worsen

Engineering Contradiction:
Improveprivacy concernsVSAvoidsimulation accuracy
Core Design Contradiction:
Object-affected harmful factorsVSMeasurement precision

Solution Approach 1:

The system changes the parameters of the data through machine learning transformations, adjusting statistical properties, movement patterns, and spatial distributions to accurately reflect real-world traffic behavior. By carefully controlling these parameter transformations, the system generates synthetic data that maintains high simulation accuracy while preserving privacy.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The system incorporates feedback mechanisms where machine learning models continuously learn from real traffic data patterns and adjust synthetic data generation accordingly. This feedback loop ensures that the synthetic data progressively improves in accuracy and realism while maintaining privacy protection, allowing the system to refine its representations over time.

Inventive Principle:
Principle #23Feedback

3Measurement precision

If machine learning models are trained on real-world data, then synthetic data realism is improved, but data processing complexity and computational resources worsen

Engineering Contradiction:
Improvesynthetic data realismVSAvoiddata processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system segments the data processing into distinct stages: collecting real-world data, training machine learning models on this data, and generating synthetic data from the trained models. This segmentation allows each stage to be optimized independently, reducing overall processing complexity by separating the computationally intensive model training from the more efficient synthetic data generation phase.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4668243A1System and method for simulating traffic using agent-based modeling
Publication Date: 2025.12.24 ALLSTATE INSURANCE COMPANY
  • EP4668243A1 patent drawingFigure 1
  • EP4668243A1 patent drawingFigure 2
  • EP4668243A1 patent drawingFigure 3

AI summary

Implementations claimed and described herein provide systems and methods for simulating traffic using synthetic data based on agent-based modeling that simulates real drivers in a particular geographic area and time frame. In one implementation, inputting, in a machine-learning model of a simulation system, real-world agent-based position data associated with a custom selection of a geographic area and a time frame. The machine-learning model of the simulation system outputs metrics associated with synthetic agent-based position data over time within a map for the simulated environment and the variation of the simulated environment, wherein the metrics represents synthetic movement behavior of agents associated with synthetic individuals based on real movement behavior associated with the geographic area and the time frame.