RAMU (Really Awesome Mobile Unit)

RAMU is a low-cost, wheel encoder-less, holonomic mobile base that is capable of autonomous navigation. RAMU is powered by ROS 2 and Navigation 2 systems. This repository contains some information on RAMU along with ROS 2 packages and launch files to get RAMU up and running.

Click the image below to watch RAMU autonomously navigate in an indoor environment.

Hardware Design

RAMU's chassis is constructed from 20x20 aluminium extrusion members which are connected together with 3D printed joints. The motor, battery, electronics and sensors mounts are also 3D printed. The CAD is open sourced via Onshape and can be accessed here.

Documentation of the build process is found here

BOM

1. Raspberry Pi 3B
2. Arduino Uno
3. DC geared motors 178RPM x4
4. Motor drivers x2
5. Mecanum wheels 96mm diameter x4
6. 2D lidar

Electrical and Wiring

RAMU's four wheels/motors follow the below numbering convention

   ^
   |
w1---w2
|     |
|     |
w3---w4

Each motor is controlled by a motor driver which receives PWM signals from the Arduino which in-turn receives motor velocities from the RPi.

The RPi thus has the Arduino and 2D lidar plugged into its USB ports. These devices must be available on /dev/ttyUSB1 and /dev/ttyUSB0 respectively. A 12V to 5V DC-DC converter steps down the 12V battery voltage and powers the RPi.

TODO Add wiring diagram

Software Setup

Arduino

RAMU's Arduino receives motor velocities in range [0,255] via serial from the RPi which is used to write PWM signals to the connected motors drivers. Expected format for incoming data:

"[w1, w2, w3, w4]"

where w1, w2, w3, w4 are the velocities of the wheels as seen in the arrangement above.

The code to be uploaded to the arduino is located at arduino/MotorDriver.

RPi

RAMU's RPi is running Ubuntu 20.04 Server. The following binaries are installed on the RPi

• ROS 2 Foxy
• Navigation 2

Other dependencies:

python3 -m pip install pyserial

Setup a ROS 2 workspace as follows

mkdir ~/ws_mecanum_bot/src -p
cd ~/ws_mecanum_bot/src
git clone https://github.com/Yadunund/mecanum_bot.git
git clone https://github.com/Yadunund/ydlidar_ros.git -b X2l-foxy-devel

Install all ROS 2 dependencies

cd ~/ws_mecanum_bot
rosdep install --from-paths src --ignore-src --rosdistro foxy -yr

Next build the workspace

cd ~/ws_mecanum_bot
source /opt/ros/foxy/setup.bash
colcon build --parallel-workers 1

Simulation

The mecanum_bot can be fully simulated with the latest Gazebo.

To launch a simple simulation world with the robot model loaded,

cd ~/ws_mecanum_bot
source install/setup.bash
ros2 launch mecanum_bot_description view_robot.launch.py

Note: Press the Play button in the bottom left corner of the Gazebo GUI to start the simulation.

The simulated robot can be teleoperated by publishing Twist messages over /cmd_vel. Lidar data is published over /scan and odometry over /odom.

Running stuff and having fun

With the Arduino code uploaded and the RPi setup, we have various launch commands to get RAMU running.

First launch the lidar node

cd ~/ws_mecanum_bot
source install/setup.bash
ros2 launch ydlidar lidar.launch.xml

Next launch the base_controller and TF2 publishers

cd ~/ws_mecanum_bot
source install/setup.bash
ros2 launch mecanum_bot_bringup mecanum_bot.launch.xml

RAMU can now be teleoperated by publishing geometry_msgs/Twist messages over /cmd_vel topic. From a second machine with a joystick plugged in,

cd ~/ws_mecanum_bot
source install/setup.bash
ros2 launch mecanum_bot_bringup joystick_teleop.launch.xml

To map an environment using Cartographer,

cd ~/ws_mecanum_bot
source install/setup.bash
ros2 launch mecanum_bot_bringup cartographer.launch.xml

Save the map files into mecanum_bot_bringup/map/ as map.pgm and map.yaml.

To navigate autonomously,

cd ~/ws_mecanum_bot
source install/setup.bash
ros2 launch mecanum_bot_bringup navigation2.launch.xml