Thi projects intention is to make a public repository where anyone can create a robot with an embedded system and hoverboard wheels Using ROS2.
- Project status
- Docker
- Hoverboard firmware hack
- Hoverboard control check
- Robot design
- Robot simulation
- Robot creation
- AI model into ROS2
- Docker ready to use it in the embedded system
- Hoverboard firmware hack
- Hoverboard control with embedded system using ROS2
- Robot design
- Robot simulation
- Create the robot
- Add own AI system for semantic segmentation into ROS2
First of all we need to prepare our docker image to use it in any device.
Dockr tips:
Commands:
- Listar todas las imágenes
docker images
- Listar todos los contenedores ejecutando
docker ps
- Listar todos los contenedores
docker ps -a
- Borrar un contenedor
docker rm CONTENEDOR_ID
- Borrar una imagen
docker rmi IMAGEN_ID
- Descargar la imagen de ROS2
docker pull docker_name:tag
- check docker images in dockerhub
- Creamos el contenedor
-
Run this command in shell
docker run --name ros2 -it \ -eDISPLAY=$DISPLAY \ --device=/dev/dri:/dev/dri \ -v$HOME/curso_ros/:/tmp/curso_ros/:rw \ -v/tmp/.X11-unix:/tmp/.X11-unix \ --env="QT_X11_NO_MITSHM=1" \ $IMAGE_ID /bin/bash
-
Dockerfile commands
- ADD
- COPY
- CMD
- ENTRYPOINT
- ENV
- FROM
- MANTAINER
- ONBUILD
- RUN
- USER
- VOLUME
- WORKDIR
The firmware hack is done in VS Code by Platfor IO using the code of EFeru/hoverboard-firmware-hack-FOC repository. Main steps to upload the firmware:
- Platform IO
Check the serial protocol of the hoverboard controlling it from the embedded system.
Robot design usig FREECAD
Using the robot design and the elements to add in ROS make the URDF and add the robot with navigation control in Gazebo.
Create the robot
Train our own model for semantic segmentation using SUN-RGBD dataset and load it to use it as a ROS node.
How to use Nav2 and PlanSys2