airo-tulip 0.3.0

Python driver for the KELO Robile platform based on KELO tulip.

AIRO Tulip

The airo-tulip package is a Python port of the KELO Tulip software with some additional functionality. You can use it to control the KELO Robile drives of a mobile platform and read out relevant sensor data from it. The structure of the code base is different from that of the original C++ implementation provided by the manufacturer.

In this README, we over the structure of the airo-tulip package and discuss some design choices that were made during the implementation.

The docs folder contains additional documentation files (in no particular order):

• kelo_setup.md on how to set up the KELO hardware and software, a prerequisite to using airo-tulip
• rerun.md on how to use Rerun running on the remote KELO CPU brick
• virtual_display.md on how to enable a virtual display for use in a VNC connection without needing to connect a display to the KELO CPU brick
• external_devices.md on how to connect to other devices mounted on the KELO, with the KELO acting as the router
• how_it_works.md on how the code base is structured and why

Even though you don't need to understand or even read all of these to use airo-tulip, we do recommend it to get a holistic overview of how the platform works and why airo-tulip is implemented the way it is.

How to use this package

There are two layers to the airo-tulip package: the hardware module that interfaces with the KELO and other attached hardware, and the api module that allows (optionally remote) users of the library to interface with this hardware over a TCP connection. The code in the hardware module is supposed to be run on the KELO itself, while the code in the api module (more specifically in api.client) can be run from any device that can access the KELO over Ethernet or Wi-Fi. While it possible to use airo-tulip without api, by using hardware.platform_driver directly, this is not recommended, and we only document the usage through the api module.

Note: when interfacing with the server from a client on a remote machine, make sure that the airo-tulip versions match on client and server, or else you may observe unexpected behaviour or crashes.

Using this package implies some hardware and software set-up on the KELO mobile platform itself, which is documented in docs/kelo_setup.md. If you simply use this library for robotics experiments, this set-up will most likely have been performed for you. Still, please read that file thoroughly before continuing. If you are the first to set up your custom KELO mobile platform, it is also recommended to read that file to get up and running quickly.

Installation

From PyPI

This package is available on PyPI and can be installed with one command:

pip install airo-tulip

Note that the package requires at least Python 3.9.

For example, using pyenv to set the local Python version to 3.9:

pyenv install 3.9 && pyenv local 3.9
python3 -m venv env
source env/bin/activate
pip install airo-tulip

Or, using conda:

conda create -n airo-tulip-env python=3.9
conda activate airo-tulip-env
pip install airo-tulip

From GitHub

If you wish to install a development version, clone the repository with git and use pip to install the airo-tulip/ package in editable mode. Note that the main branch is the active development branch: you may want to check out a certain commit associated with a version tag.

Using pyenv:

git clone https://github.com/airo-ugent/airo_kelo
cd airo_kelo
pyenv install 3.9 && pyenv local 3.9
python3 -m venv env
source env/bin/activate
pip install -e airo-tulip/

Or using conda:

git clone https://github.com/airo-ugent/airo_kelo
cd airo_kelo
conda create -n airo-tulip-env python=3.9
conda activate airo-tulip-env
pip install -e airo-tulip/

Running airo-tulip on the KELO

airo_tulip.api.server provides a class TulipServer which initializes the KELO platform and accepts an incoming connection from a airo_tulip.api.client.KELORobile. To accept incoming connections from any device on the network, listen on the IPv4 address 0.0.0.0. The RobotConfiguration that must be supplied, is based on how the KELO bricks are mounted. This information is received from KELO robotics together with your platform and is specific to your use case. The EtherCAT device is also specific to your platform set-up.

To run the server, start it from a Python script:

from airo_tulip.api.server import TulipServer, RobotConfiguration
from airo_tulip.hardware.structs import WheelConfig

def create_wheel_configs():
    wheel_configs = []

    wc0 = WheelConfig(
        ethercat_number=3,
        x=0.233,
        y=0.1165,
        a=1.57
    )
    wheel_configs.append(wc0)

    wc1 = WheelConfig(
        ethercat_number=5,
        x=0.233,
        y=-0.1165,
        a=1.57
    )
    wheel_configs.append(wc1)

    wc2 = WheelConfig(
        ethercat_number=7,
        x=-0.233,
        y=-0.1165,
        a=-1.57
    )
    wheel_configs.append(wc2)

    wc3 = WheelConfig(
        ethercat_number=9,
        x=-0.233,
        y=0.1165,
        a=1.57
    )
    wheel_configs.append(wc3)

    return wheel_configs

# These values are specific to your platform!
device = "eno1"
wheel_configs = create_wheel_configs()

server = TulipServer(RobotConfiguration(device, wheel_configs), "0.0.0.0")
server.run()

Connecting to the airo-tulip server

Once you have started the server on the KELO, you can connect to it with an api.client.KELORobile instance:

from airo_tulip.api.client import KELORobile

kelo_ip = "10.10.129.21"
client = KELORobile(kelo_ip)

You can then send commands to the KELO platform by calling the methods on the client object, e.g.,

client.set_platform_velocity_target(0.5, 0.0, 0.0, timeout=1.0)

to drive approximately 0.5 meters, at 0.5 meters per second, along the platform's +X axis.

Mounted devices

Without mounting external devices on the KELO, you can pretty much only drive around (which is cool, but not very useful). Refer to docs/external_devices.md for information on how to set up and access external devices such as a UR cobot with a Robotiq gripper.

Structure

For more information on how this code base is structured, and why, please refer to docs/how_it_works.md.