A VR-driven teleop loop that streams device poses into the simulator over ROS 2, plus device drivers and a data-recording pipeline that turns recorded rosbags into HDF5 episodes.
License: Mozilla Public License Version 2.0
Support joystick control for robot waist, left/right end effector and base movement.
| No. | Function |
|---|---|
| 1 | move base of the robot |
| 2 | control left gripper |
| 3 | backtracking action |
| 4 | reset left arm and right arm |
| 5 | enable left arm pose tracking |
| 6 | start recording |
| 7 | control the waist of the robot |
| 8 | control right gripper |
| 9 | reset body and head |
| 10 | enable right arm pose tracking |
- Connect to the same LAN as the computer.
- Start AIDEA Vision App in resource library.
- Choose Wireless Connection and enter the IP of the computer.
Note: The AIDEA Vision App has been uploaded; you can access it here: https://modelscope.cn/datasets/agibot_world/GenieSim3.0-Dataset/tree/master/app . APK installation for the Pico Enterprise Edition HMD can be performed via a computer. When the HMD is connected via USB, its internal storage becomes accessible as a standard drive on the computer. The APK file can then be installed by copying it directly to this drive.
Execute the following command under genie_sim root directory outside the docker
container (after geniesim docker5.1 up succeeds):
genie_sim$ ./source/geniesim_teleop/scripts/autoteleop.shIf you want to change tasks, please modify the task_name and sub_task_name
fields in ./source/geniesim/config/teleop.yaml.
Execute the following command under genie_sim root directory outside the docker
container:
genie_sim$ ./source/geniesim_teleop/scripts/autoteleop_post_process.sh {SUB_TASK_NAME}
# for example
genie_sim$ ./source/geniesim_teleop/scripts/autoteleop_post_process.sh open_doorNote: To improve collection efficiency, data collection and data post-processing can be run in parallel.