https://github.com/wkentaro/safepicking

SafePicking: Learning Safe Object Extraction via Object-Level Mapping, ICRA 2022
https://github.com/wkentaro/safepicking

artificial-intelligence computer-vision deep-learning machine-learning reinforcement-learning robotics ros

Last synced: 7 days ago
JSON representation

SafePicking: Learning Safe Object Extraction via Object-Level Mapping, ICRA 2022

• Host: GitHub
• URL: https://github.com/wkentaro/safepicking
• Owner: wkentaro
• License: other
• Created: 2022-02-10T15:49:21.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2022-03-16T12:29:01.000Z (over 2 years ago)
• Last Synced: 2024-11-01T10:35:11.607Z (14 days ago)
• Topics: artificial-intelligence, computer-vision, deep-learning, machine-learning, reinforcement-learning, robotics, ros
• Language: Python
• Homepage: https://safepicking.wkentaro.com
• Size: 38.5 MB
• Stars: 53
• Watchers: 4
• Forks: 2
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
SafePicking

Learning Safe Object Extraction via Object-Level Mapping




  Kentaro Wada,

  Stephen James,

  Andrew J. Davison

  


  Dyson Robotics Laboratory,

  Imperial College London

  


  IEEE International Conference on Robotics and Automation (ICRA), 2022





  Installation |

  Usage |

  Paper |

  Video |

  Webpage





  



---



  

  

  



## Installation

### Python project only

```bash

make install

source .anaconda3/bin/activate

./checks/check_motion_planning.py

```

### ROS project

- `robot-agent`: A computer with CUDA and a GPU installed for visual processing.

- `robot-node`: A computer with a real-time OS for a Panda robot.

#### @robot-agent

```bash

make install

source .anaconda3/bin/activate

cd ros/

make install

source devel/setup.sh

```

#### @robot-node

```

cd ros/

source /opt/ros/noetic/setup.sh

catkin build morefusion_panda

rosrun morefusion_panda create_udev_rules.sh

catkin build safepicking_ros

```

## Usage

### Training & Inference

```bash

cd examples/picking/

# download pile files at ~/.cache/safepicking/pile_generation/

./download_piles.py

./train.py --model fusion_net --noise

./learned.py logs/XXX/weights/YYY

# use pretrained model

./download_pretrained_models.py

# inference in the test environments: ~/.cache/safepicking/pile_generation/00009000 - 00009999.pkl

./learned.py --weight-dir logs/20210709_005731-fusion_net-noise/weights/84500 \

             ~/.cache/safepicking/pile_generation/00009000.pkl

```



  

  

  

  

  



### Robotic demonstration

```bash

robot-node  $ roslaunch safepicking_ros panda_control.launch

robot-agent $ roslaunch safepicking_ros setup.launch

robot-agent $ rosrun safepicking_ros safepicking_task_interface.py

>>> self.run([YcbObject.PITCHER])

```



## Citation

```

@inproceedings{Wada:etal:ICRA2022a,

  title={{SafePicking}: Learning Safe Object Extraction via Object-Level Mapping},

  author={Kentaro Wada and Stephen James and Andrew J. Davison},

  booktitle={IEEE International Conference on Robotics and Automation (ICRA)},

  year={2022},

}

```