XinyangJiang / rmp2

Code for R:SS 2021 paper RMP2: A Structured Composable Policy Class for Robot Learning.

RMP2

Code for R:SS 2021 paper RMP2: A Structured Composable Policy Class for Robot Learning. [Paper]

Installation

git clone https://github.com/UWRobotLearning/rmp2.git
cd rmp2
conda env create -f environment.yml
. startup.sh

Hand-designed RMP2 for Robot Control

To run a goal reaching task for a Franka robot:

python examples/rmp2/rmp2_franka.py

To run a goal reaching task for a 3-link robot:

python examples/rmp2/rmp2_3link.py

Training RMP2 Policies with RL

Note: The instruction below is for the 3-link robot. To run experiments with the franka robot, simply replace 3link by franka.

To train an NN policy from scratch (without RMP2):

python run_3link_nn.py

To train an NN residual policy:

python run_3link_nn.py --env 3link_residual

To train an RMP residual policy:

python run_3link_residual_rmp.py

To restore training of a policy:

python restore_training.py --ckpt-path ~/ray_results/[EXPERIMENT_NAME]/[RUN_NAME]/

To visualize the trained policy:

python examples/rl/run_policy_rollouts.py --ckpt-path ~/ray_results/[EXPERIMENT_NAME]/[RUN_NAME]/

Citation

If you use this source code, please cite the below article,

@inproceedings{Li-RSS-21,
    author = "Li, Anqi and Cheng, Ching-An and Rana, M Asif and Xie, Man and Van Wyk, Karl and Ratliff, Nathan and Boots, Byron",
    booktitle = "Robotics: Science and Systems ({R:SS})",
    title = "{{RMP}2: A Structured Composable Policy Class for Robot Learning}",
    year = "2021"
}