teleop_data_collect.md

Running the teleop

For Robots

To use the OpenTeach teleop module with robots, open the VR Application in your Oculus Headset. On the robot server side, start the controllers first for robots followed by the following command to start the teleop :

Franka and Kinova mirrors human arm motions and retargets it to the robot.

Allegro Hand

python teleop.py robot=allegro

Bimanual XARM

python teleop.py robot=bimanual

Franka

python teleop.py robot=franka

Kinova

python teleop.py robot=kinova

Xela Hand

python teleop.py robot=xela_hand run_xela=True

Allegro Franka

python teleop.py robot=allegro_franka

Xela Hand Franka

python teleop.py robot=xela_hand_franka run_xela=True

Allegro Kinova

python teleop.py robot=allegro_kinova

For Simulation

To use the Open-Teach teleop module, open the VR Application in your Oculus Headset. On the server side, use the following command to start the teleop :

Allegro Sim

python teleop.py robot=allegro_sim sim_env=True

Libero Sim

python teleop.py robot=libero_sim sim_env=True

Data Collection

The Data Collection module saves the robot states , cameras sensors output as a video and sensor states.

For Robots

Without Xela Sensors

python data_collect.py robot=allegro/kinova/franka/allegro_franka/allegro_kinova/bimanual demo_num=1

With Xela Sensors

python data_collect.py robot=xela_hand_franka demo_num=1 is_xela=True

For Simulation

python3 data_collect.py robot=allegro_sim/libero_sim demo_num=1 sim_env=True

Note: Remember to enter your network IP on the server config

The data saves camera stream in the optimized form of .avi videos and saves depth and robot information in the form of h5 files.