This repository provides a customizable SpaceX rocket landing simulation environment for OpenAI Gym, targeting reinforcement learning researchers and developers. It offers a challenging physics-based task with both discrete and continuous control inputs, allowing for flexible agent development and evaluation.

How It Works

The environment simulates rocket physics using the Box2D library, modeling factors like gravity, thrust, and fuel consumption. It supports configurable state representations, including velocities, and offers distinct discrete and continuous action spaces for agent interaction. This dual-input approach allows for experimentation with different control strategies.

Quick Start & Requirements

• Primary install: pip install -e . within the cloned repository directory.
• Prerequisites: OpenAI Gym.
• Usage: env = gym.make("gym_rocketlander:rocketlander-v0")

Highlighted Details

• Customizable state variables including position, angle, leg contact, throttle, and gimbal.
• Optional inclusion of velocity states (x, y, angular).
• Supports both discrete and continuous control inputs for agent interaction.

Maintenance & Community

No information on maintainers, community channels, or roadmap is available in the README.

Licensing & Compatibility

The README does not specify a license. Compatibility for commercial use or closed-source linking is undetermined.

Limitations & Caveats

The README lacks details on specific Box2D version requirements, potential performance bottlenecks, or known issues. The absence of a specified license and community information may hinder adoption.