Quick Start Guide
Contents
Quick Start Guide#
Here we provide a list of functionalities and their corresponding documentation link.
Here is a video (better viewed when zoom in) that shows the easy usage of Deoxys.
` And we provide a table quick-start links for users to follow, find their needs, and get up to speed.
Functionatlities | Link |
---|---|
How to turn on/off the Franka Emika Panda | running-robots page |
How should I install the real-time kernel Linux system? | system prerequisite page |
How should I install Deoxys codebase? |
codebase installation page |
Use a spacemouse to control the robot | using-teleoperation-devices page |
How should I use logger in Deoxys ? |
logger page |
How should I use the controllers in Deoxys ? |
using-robot-controllers page |
How should I replay a sequence teleoperated motion? | record-and-replay page |
How should I design some hand-crafted motor programs? | handcrafting-program page |
Easy usability#
We try our best to design the easy-to-use api to facilitate research. Here we quickly go through the minimal effort you are going to need for your own project once installation and prerequisites are met (see the table above for reference).
Initialize python interface#
from deoxys.franka_interface import FrankaInterface
robot_interface = FrankaInterface("config\/charmander.yml")
controller_type = "OSC_POSE"
from deoxys.utils.config_utils import get_default_controller_config
controller_cfg = get_default_controller_config(controller_type=controller_type)
Launch robot arm and gripper script#
$ ./auto_scripts/auto_arm.sh config/charmander.yml
$ ./auto_scripts/auto_gripper.sh config/charmander.yml
Stream your sensory data and control the robot#
This codebase does not assume the streaming source of sensory data. In this example usage, we assume we got RGB data in images
variable
for _ in range(num_steps):
imgs = ... # Get your sensory data like RGB images
action = model(imgs)
robot_interface.control(
controller_type=controller_type,
action=action,
controller_cfg=controller_cfg
)