Scope of the initiative

Still under construction!

The vision of open traffic science is to bridge the world and times for traffic science. The initiative is to open up the knowledge (fundamental problems) of traffic science to everyone, followed with open data, open tools, open tutorials, and open discussions.

Background

The origins of traffic science date back to 1950s, when pioneers like Leslie C. Edie, Robert Herman, Robert E. Chandler, Denos Gazis, Elliott W. Montroll, Refrey B. Potts, and Richard W. Rothery paved the way for this field. They conducted extensive field experiments, initiated the ISTTT symposium series, and authored several influential textbooks.

Challenges with the times

Since the rise of big data and deep learning (starting around 2015), along with advancements in reinforcement learning (notably since AlphaGo) and large models (from 2022 onwards), transportation and traffic research has undergone significant transformations in various aspects.

Challenges with the world

There are couples of labs in the world leading the efforts to turn traffic science into the next level. However, the knowledge is still not open to everyone. Or saying, not all the students can have access to the same level of resources as the top labs. For example, NGSIM data, nearly all the traffic researchers know it and can use it. But what about the ring experiment data in 2006 in Japan? It actually exists, but not everyone knows it.

Scope

The problems of transportation and traffic are very old ones.

In the era of connected and automated vehicles and artificial intelligence,

We plan to revisit traffic science with emerging technologies and open data.

We need to identify the fundamental questions in our research community. No matter how the technologies evolve, the fundamental questions remain consistent. The philosophy behind this initiative is to mentor the younger generation using first principle.

The first principle is a foundational concept or assumption that cannot be deduced from any other proposition and serves as the basic building block for understanding and reasoning.

For people with diverse backgrounds and purposes, we aim to create a “knowledge tree” that provides various learning paths, enabling personalized learning experiences.

As a textbook-style initiative, we would pay more attention to the scientific part, although we might be able to include more engineering details. Or, we should separate those tedious (although important) data processing with true science and knowledge. This is why we recognize ourselves as traffic scientists.

Plan

At the initial stage, Xingmin and Junyi could start from writing the main content at the beginning. Getting some senior advising professors to review the contents. Xingmin is familiar with urban traffic while Junyi is more familiar with freeway traffic. Topics like travel behavior and traffic planning are beyond the scope of the first iteration.

Scale up: once we have certain scale, we can start to advise and open to public. I think it might be a good idea to make it open in twofold: 1) accessible to everyone. 2) everyone can also contribute to it.

Open Science: Making science more accessible, inclusive and equitable for the benefit of all.

Core value: Open knowledge to everyone, learn & contribute (issues & pull request)

Stage 1: online book and blogging

Stage 2: making it more iterative, like open-source community

Main content

Online blogging (lecture)

  1. Traffic measurement
  2. Traffic model
    • First-order continuum traffic flow model (LWR model)
    • Numerical solutions of LWR model, Godunov scheme, CTM
  3. Signalized intersections
    • Scientific merit in traffic signal control (what a researcher should know from MUCTD)
    • Probabilistic traffic flow model for signalized intersections
  4. Data platform for connected vehicles trajectories in large-scale networks

Online coding examples (Colab)

Course project (large-scale examples)


(Posted by Junyi Ji and Xingmin Wang on October 1, 2024.)