Farama-Foundation/HighwayEnv

A collection of environments for autonomous driving and tactical decision-making tasks

What it solves

It provides a standardized set of simulation environments for developing and testing autonomous driving agents. It focuses on tactical decision-making tasks, allowing researchers and developers to train AI agents to handle complex road scenarios without needing a full-scale physical simulator.

How it works

Built on the Gymnasium (formerly OpenAI Gym) interface, the project provides a collection of driving scenarios. Users can define the environment configuration (such as the number of lanes) and implement their own AI policies to control the vehicle. The environment handles the physics and rules of the road, returning observations and rewards to the agent based on its actions.

Who it’s for

It is designed for researchers and developers working on autonomous driving, reinforcement learning, and tactical decision-making in transportation.

Highlights

  • Diverse Scenarios: Includes 10 distinct driving families, such as intersections, roundabouts, parking, and highway merging.
  • Flexible Variants: Offers specialized versions of environments, including multi-agent, continuous-control, and connected-lane variants.
  • Gymnasium Integration: Fully compatible with the Gymnasium API, making it easy to integrate with popular RL libraries like Stable Baselines3.
  • Broad Support: Compatible with Linux, macOS, and Windows.

Related

  • Project
    Farama-Foundation/Gymnasium

    A standard Python API and environment suite for developing and comparing reinforcement learning algorithms.

  • Project
    ServiceNow/BrowserGym

    An open-source framework that provides a unified interface for researching and evaluating web agents across multiple benchmarks and open-ended web tasks.

  • Project
    Farama-Foundation/MiniWoB-plusplus

    A collection of over 100 web interaction environments for training and AI agents to interact with web browsers using the Gymnasium API and Selenium.

  • Project
    Farama-Foundation/Minigrid

    A collection of lightweight, discrete grid-world environments for Reinforcement Learning research, supporting goal-oriented tasks and grounded language learning.

  • Project
    Farama-Foundation/Stable-Retro

    A maintained fork of gym-retro that turns classic video games into Gymnasium environments for reinforcement learning training and benchmarking.