Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Data-Driven Traffic Simulation for an Intersection in a Metropolis

Chengbo Zang, Mehmet Kerem Turkcan, Gil Zussman, Javad Ghaderi, Zoran Kostic

TL;DR

The paper tackles realistic micro-level traffic simulation at urban intersections by learning from real-world trajectory data. It proposes a two-stage approach: first, generate coarse priors by sampling from time- and space-conditioned Gaussian Mixture Models over entry/exit states, speeds, and way-points; second, refine these priors with TrajNet++ trajectory forecasting using goal or waypoint supervision in an iterative prediction loop. The key contributions include a full data-to-simulation pipeline—data collection, statistical modeling, prior generation, and deep refinement—that enables both autonomous and human-controlled simulation, with a peak result of $0.36$ m FDE at $20$ FPS on an NVIDIA A100. This framework supports realistic multi-agent interactions and could enhance collision alerts, traffic management, and digital twin applications, though broader geographic validation across multiple intersections remains for future work.

Abstract

We present a novel data-driven simulation environment for modeling traffic in metropolitan street intersections. Using real-world tracking data collected over an extended period of time, we train trajectory forecasting models to learn agent interactions and environmental constraints that are difficult to capture conventionally. Trajectories of new agents are first coarsely generated by sampling from the spatial and temporal generative distributions, then refined using state-of-the-art trajectory forecasting models. The simulation can run either autonomously, or under explicit human control conditioned on the generative distributions. We present the experiments for a variety of model configurations. Under an iterative prediction scheme, the way-point-supervised TrajNet++ model obtained 0.36 Final Displacement Error (FDE) in 20 FPS on an NVIDIA A100 GPU.

Data-Driven Traffic Simulation for an Intersection in a Metropolis

TL;DR

The paper tackles realistic micro-level traffic simulation at urban intersections by learning from real-world trajectory data. It proposes a two-stage approach: first, generate coarse priors by sampling from time- and space-conditioned Gaussian Mixture Models over entry/exit states, speeds, and way-points; second, refine these priors with TrajNet++ trajectory forecasting using goal or waypoint supervision in an iterative prediction loop. The key contributions include a full data-to-simulation pipeline—data collection, statistical modeling, prior generation, and deep refinement—that enables both autonomous and human-controlled simulation, with a peak result of m FDE at FPS on an NVIDIA A100. This framework supports realistic multi-agent interactions and could enhance collision alerts, traffic management, and digital twin applications, though broader geographic validation across multiple intersections remains for future work.

Abstract

We present a novel data-driven simulation environment for modeling traffic in metropolitan street intersections. Using real-world tracking data collected over an extended period of time, we train trajectory forecasting models to learn agent interactions and environmental constraints that are difficult to capture conventionally. Trajectories of new agents are first coarsely generated by sampling from the spatial and temporal generative distributions, then refined using state-of-the-art trajectory forecasting models. The simulation can run either autonomously, or under explicit human control conditioned on the generative distributions. We present the experiments for a variety of model configurations. Under an iterative prediction scheme, the way-point-supervised TrajNet++ model obtained 0.36 Final Displacement Error (FDE) in 20 FPS on an NVIDIA A100 GPU.
Paper Structure (16 sections, 5 equations, 3 figures, 1 table, 2 algorithms)

This paper contains 16 sections, 5 equations, 3 figures, 1 table, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Method
  4. Data Collection
  5. Statistical Analysis
  6. Generation of Prior Trajectories
  7. Deep-Learning-Based Trajectory Refinement
  8. Simulation Algorithm
  9. Experiments
  10. Dataset and Evaluations
  11. Controlled Simulation
  12. Autonomous Simulation
  13. Simulation Quality
  14. Outliers
  15. Model Configuration
  16. ...and 1 more sections

Figures (3)

  • Figure 1: Overall workflow of agent generation.(a) Real-world trajectories collected from the intersection (vehicles in purple and pedestrians in orange). (b) Examples of different types of trajectories categorized by GMMs. (c) Coarse way-points sampled from GMMs and interpolated prior trajectories. (d) Final trajectories refined by deep forecasting models compared to the coarsely sampled prior trajectories in (c).
  • Figure 2: Distribution of agent densities over $24$ hours. The $x$-axis is the ToD and the $y$-axis is the hourly average pedestrian and vehicle counts.
  • Figure 3: Simulation results and experiments.(a) - (b) Controlled simulation of a potential collision and the reaction of the trajectory forecasting model. (c) Outliers identified in pedestrians by thresholding the likelihood of the trajectories. (d) - (f) Comparison of different supervision schemes for TrajNet++. Under the same priors, different results of refined trajectories are given by (d) no supervision, (e) final destination as supervision, and (f) way-points iteratively sampled from the priors as supervision.