urdfenvs 0.4.0

Simple simulation environment for robots, based on the urdf files.

Generic URDF robots

Generic URDF robots

In this package, generic urdf robots and a panda gym environment are available. The goal is to make this environment as easy as possible to deploy. Although, we used the OpenAI-Gym framing, these environments are not necessarly restricted to Reinforcement-Learning but rather to local motion planning in general.

Point Robot	Point Robot with Keyboard Input	Non-Holonomic Robot

Tiago Robot	Tiago Robot with Keyboard Input

Panda Robot	Albert Robot

Getting Started

This is the guide to quickle get going with urdf gym environments.

Pre-requisites

• Python >3.6, <3.10
• pip3
• git

Installation

You first have to downlad the repository

git clone [e-mail]:maxspahn/gym_envs_urdf.git

Then, you can install the package using pip as:

pip3 install .

Optional: Installation with poetry

If you want to use poetry, you have to install it first. See their webpage for instructions docs. Once poetry is installed, you can install the virtual environment with the following commands. Note that during the first installation poetry update takes up to 300 secs.

poetry install

The virtual environment is entered by

poetry shell

Inside the virtual environment you can access all the examples.

Examples

Run example

You find several python scripts in examples/. You can test those examples using the following (if you use poetry, make sure to enter the virtual environment first with poetry shell)

python3 point_robot.py

Replace point_robot.py with the name of the script you want to run.

Use environments

In the examples, you will find individual examples for all implemented robots. Environments can be created using the normal gym syntax. Gym environments rely mostly on three functions

• gym.make(...) to create the environment,
• gym.reset(...) to reset the environment,
• gym.step(action) to step one time step in the environment.

For example, in examples/point_robot.py, you can find the following syntax to make, reset and step the environment.

env = gym.make('pointRobotUrdf-vel-v0', dt=0.05, render=True)
ob = env.reset(pos=pos0, vel=vel0)
ob, reward, done, info = env.step(action)

The id-tag in the make command specifies the robot and the control type. You can get a full list of all available environments using

from gym import envs
print(envs.registry.all())

Go ahead and explore all the examples you can find there.