Collective Intelligence for 2D push Manipulations with Mobile Robots

So Kuroki, Tatsuya Matsushima, Jumpei Arima, Hiroki Furuta, Yutaka Matsuo, Shixiang Shane Gu, and Yujin Tang

The University of Tokyo, Japan, Matsuo Institute, Japan, Google Research, Brain Team
IEEE Robotics and Automation Letters (RA-L) with IROS
[Project page] [PDF]

generating experts

For gradient base difftaichi

CUDA_VISIBLE_DEVICES=0 python3 -m plb.algorithms.solve --algo action --env_name multi_bc_rope_6-v1

creating observation

python3 data_create.py

train policy

CUDA_VISIBLE_DEVICES=0,1 python3 scripts/train_bc_transformer_nearest.py --env_name multi_bc_rope_6-v1
CUDA_VISIBLE_DEVICES=0,1 python3 scripts/train_bc_transformer_nearest_random_delay.py --env_name multi_bc_rope_6-v1
CUDA_VISIBLE_DEVICES=0,1 python3 scripts/train_bc_mlp.py --env_name multi_bc_rope_6-v1

test

CUDA_VISIBLE_DEVICES=0,1 python3 -m plb.algorithms.solve_bc_transformer_nearest --algo action --env_name multi_bc_rope_6-v1
CUDA_VISIBLE_DEVICES=0,1 python3 -m plb.algorithms.solve_bc_transformer_nearest_random_delay --env_name multi_bc_rope_6-v1
CUDA_VISIBLE_DEVICES=0,1 python3 -m plb.algorithms.solve_bc_mlp --env_name multi_bc_rope_6-v1

test in real-world

rosrun roomba_control solve.py --env_name multi_bc_rope_{n}-v1 --real_demo

Acknowledgements

Our physics simulation is based on PlasticineLab.
Our ros system refer to roomba hack.