Franka Kitchen

Multitask environment in which a 9-DoF Franka robot is placed in a kitchen containing several common household items. The goal of each task is to interact with the items in order to reach a desired goal configuration.

The tasks can be selected when the environment is initialized passing a list of tasks to the tasks_to_complete argument as follows:


import gymnasium as gym
import gymnasium_robotics

gym.register_envs(gymnasium_robotics)

env = gym.make('FrankaKitchen-v1', tasks_to_complete=['microwave', 'kettle'])

The possible tasks to complete are:

Task

Description

bottom burner

twist control knob to activate bottom left burner in the stove.

top burner

twist control knob to activate top left burner in the stove.

light switch

move a lever switch to turn on a light over the burners.

slide cabinet

slide open the cabinet door.

hinge cabinet

open a hinge cabinet door.

microwave

open the microwave door.

kettle

move the kettle from the bottom burner to the top burner.

References

These environments were first introduced in “Relay Policy Learning: Solving Long-Horizon Tasks via Imitation and Reinforcement Learning” by Abhishek Gupta, Vikash Kumar, Corey Lynch, Sergey Levine, Karol Hausman, and later modified in “D4RL: Datasets for Deep Data-Driven Reinforcement Learning” by Justin Fu, Aviral Kumar, Ofir Nachum, George Tucker, Sergey Levine. Both publications can be cited as follows:

@article{gupta2019relay,
  title={Relay policy learning: Solving long-horizon tasks via imitation and reinforcement learning},
  author={Gupta, Abhishek and Kumar, Vikash and Lynch, Corey and Levine, Sergey and Hausman, Karol},
  journal={arXiv preprint arXiv:1910.11956},
  year={2019}
}
@misc{fu2020d4rl,
    title={D4RL: Datasets for Deep Data-Driven Reinforcement Learning},
    author={Justin Fu and Aviral Kumar and Ofir Nachum and George Tucker and Sergey Levine},
    year={2020},
    eprint={2004.07219},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}