Cerebrum Documentation

Repository Part of Roboost Project

This repository, a crucial segment of the Roboost project, hosts high-level functionalities of the Roboost robot in ROS2 packages. These packages are tailored for deployment on the robot's onboard computer using Docker and allow local development through a VSCode devcontainer.

Project Information:

• Roboost Project Website
• Roboost GitHub Organization

Description

The Cerebrum repository incorporates ROS2 packages essential for the operation and control of the Roboost robot. These packages integrate to manage navigation, vision processing, and other more advanced functionalities.

Here you can see the simulated robot in Gazebo:

And here the integration with slam-toolbox:

Packages

roboost

Core functionalities of the robot, including control, vision, and navigation.

Launch Files

• camera.launch.py: Launches the camera node.

  ros2 launch roboost camera.launch.py

• joy_control.launch.py: Activates joystick control.

  ros2 launch roboost joy_control.launch.py

• mecanum_sim.launch.py: Simulates the robot in Gazebo.

  ros2 launch roboost mecanum_sim.launch.py world:=src/roboost/worlds/home.world

• mecanum_tf_broadcast.launch.py: Publishes the robot's TF tree and URDF model.

  ros2 launch tf_broadcast_package mecanum_tf_broadcast_launch.py

yolov8_ros

Integrates YOLOv8 for object detection.

Test Usage

• Camera stream:

  ros2 launch roboost camera.launch.py

• YOLOv8 node:

  ros2 launch yolov8_bringup yolov8.launch.py device:=cpu input_image_topic:=/image_raw

slam-toolbox

For mapping and localization using SLAM.

Launching

• Start mapping:

  ros2 launch slam_toolbox online_async_launch.py params_file:=./src/roboost/config/mapper_params_online_async.yaml

Installation

TODO: Add more detailed instructions, including config and launch files.

Prerequisites

• VSCode
• Docker

Building and Deployment

• Clone the repository.
• Open in VSCode.
• Use Docker for building and deployment:
  • Build the Docker image:

docker build -t roboost-cerebrum .

• Run a container from the image:

docker run -it --rm --name cerebrum_container roboost-cerebrum

• Inside the container, navigate to the workspace and build the ROS2 packages:

colcon build
source install/setup.bash

Launch the desired ROS2 nodes using the launch files provided in the documentation.

Usage with Roboost-Mecanum Robot

Start Micro-ROS Agent

docker run --rm -it --device=/dev/ttyUSB0 --net=host microros/micro-ros-agent:humble serial --dev /dev/ttyUSB0

or for multiple devices:

sudo docker run -it --rm -v /dev:/dev --privileged --net=host microros/micro-ros-agent:humble multiserial --devs "/dev/ttyUSB0 /dev/ttyUSB1"

The agent should now be running and waiting for a client to connect. Now, the other nodes can be launched as usual.

Additional Components (Planned)

• ChatGPT integration
• Speech recognition and synthesis
• NAV2 stack implementation