Simple-ROS-Arduino-odometry

I started this project because I'm done with rosserial: it is just too bulky for the arduino's 2KB flash, and almost impossible to debug. The only two things it does is refuse to connect or report trying to sync. Trust me, it's much quicker to just write your own serial communication. It's quicker even I've never written any Python and actually learning Python at the same time!
For more information, please see About the Code section below.

PLEASE CHECK ENCODER SECTION BELOW BEFORE PROCEEDING

Getting Started

Prerequisites

Tested and built with ROS-kinetic, Python 2.7, with Ubuntu 16.04LTS, but should work with any version near enough.
Please make sure you have the following items set-up or at least you know what you're doing:

• install ROS (tested in ROS-kinetic)
• setup ROS catkin workspace (it is recommanded to source the setup in .bashrc)
• Install the Arduino IDE ( I recommand you to put the Arduino IDE under /opt/ before running install.sh )
• Add your user to dialout group as required by the Arduino IDE (Hint1:sudo usermod -a -G dialout yourUserName, Hint2: you have to login again, just reboot!)
• (For non-quadrature encoders aka tachometers) We need to subscribe to topic /cmd_vel(geometry_msgs.msg::Twist) to determine the direction of the wheel rotation.

Installation

1. Go to your catkin directory, clone the project into folder /serial_odom along with all the submodules(Arduino library).

   cd ~/catkin_ws/src
   git clone --recurse-submodules -j8  https://github.com/EwingKang/Simple-ROS-Arduino-odometry serial_odom
   

2. (Optional) Link(or move) the Ardunio program to workspace. It is possible to open the (project)/arduino_files/serial_odom/serial_odom.ino with Arduino IDE directly. But I find it faster to find if everything is under Arduino IDE directory. You'll see a link folder in the Arduino/ afterwards

   ln -s ~/catkin_ws/src/serial_odom/arduino_files/serial_odom/ ~/Arduino/
   

3. "Compile" the file with catkin_make so ros recgonize the package

   cd ~/catkin_ws/
   catkin_make
   

4. Upload the Arduino code with Arduino IDE Please setup your vehicle parameter first!

Vehicle settings

To make sure you have the correct result, it is necessary to setup parameters matching your vehicle geometry. Since odometry is calculated "on-board" the Arduino, you should change the config header accordingly

• Physical parameters in vehicle_config.h:

  #define ENC_TYPE 4
  #define ENC_ENABLE_CUMU		// enable cumulative counter
  
  //========== VEHICLE CONFIG ==========//
  #define WHEEL_RAD 0.033f		// wheel radius in m
  #define B 0.1f			// b = wheel separation distance in m
  #define ENC_REDUCTION 21		// encoder reduction ratio = (wheel rpm)/(encoder rpm)
  #define MAX_WHEEL_RPS ((float)201/(float)60)	// maximum wheel revolution per second
  
  //========== ENCODER CONFIG ==========//
  #define ENC_TPR 334			// encoder tooth count
  #define WHEEL_PPR (ENC_TYPE*ENC_TPR*ENC_REDUCTION)	// pulse count per wheel revolution
  

• Special (cheap/homemade) Encoders: Please check the Tachometer section below.

Run the node

Hook up all your cable, make sure you have the correct device (in this case /dev/ttyACM0), and simply rosrun:

rosrun serial_odom serial_odom.py _port:=/dev/ttyACM0

The code will reset the Arduino automatically, and you'll see:

1. Node starting
2. Independent communication thread starting
3. Rebooting Arduino
4. Arduino standby message
5. Signal to start Arduino
6. First packet received and /odom is now published

Check the result

In another terminal, you can check the result by:

• rostopic You be able to see the /odom topic with rostopic list. Use rostopic echo /odom to see the raw data
• rviz: Run rviz and the ROS visualization tool will begin. Set the frame to "Odom" and you'll see "base_link" moving around.

ROS node parameters

All the parameters are under namespace /serial_odom

• ~odom_topic (default: 'odom')

• ~base_id (default: 'base_link') # base frame id

• ~odom_id (default: 'odom') # odom frame id

• ~enable_tf (default: True)

• ~port (default: '/dev/ttyACM0') # port

• ~baudrate (default: 115200) ) # baudrate

• ~serial_timeout (default: 10) )

• ~odom_freq (default: 100) ) # hz of communication*

• ~tx_freq (default: 5) ) # hz of communication

• ~quadrature (default: True)

• ~cmd_vel (default: 'cmd_vel')

• ~cmd_vel_timeout (default: 3) ) #

Project details

TODO

Tachometer

TODO

Encoder

I'm using a cheap ass 175NTD(~5$) chinese-made encoder like this one and it is super lame. Despite the claimed AB phase output with a 100 grid encoder plate, it just simply connot hangle that. And I have to make my own 25 grid encoder plate out of catdboard. Anyway, you're welcome to take this code and modified it to adopt proper A/B phase encoder, or even better, you can donate me some proper encoders and I'll change the code for you. Currectly, the direction information comes from a ROS topic, which piggyback the motor controller.

Python

I'm using Python 2.7 that comes with Ubuntu-16.04 & ROS-kinetic, with following modules:

• rospy and its peripherals
• rostf
• sys, time, math
• serial (pySerial)
• threading

pySerial is used to easily handle the serial port and capture the data. Serial communication is handled in a seperate thread wih Python module "threading", which make it very elastic, powerful, and stable.

Arduino

On the Arduino side, data is struted and can be easily modified. [0x00 0xFF] is used as the starting signal of each packet.

Line-by-line explaination

TODO: see wiki

Other costumizable parameters

• Arduino sceduling system

  #define SERIAL_PERD 100000  // Serial tx/rx check period, 100 ms is 10 Hz
  #define CAR_PERD 20000	    // Main calculation period, 20 ms is 50 Hz
  

• ROS node (Python script) make sure your

  self.comm_freq = float( rospy.get_param('~odom_freq', '15') ) # hz of communication
  

  is higher than the rate the Arduino sends the message. For example:

  ***THIS IS GOOD***
  #define SERIAL_PERD 100000  // Serial tx/rx check period, 100 ms is 10 Hz
  ---------
  self.comm_freq = float( rospy.get_param('~odom_freq', '15') ) # hz of communication
  

  If you recieve less frequently than the transmit rate, packet will be lost and you'll potentially have a jumping odometry. Equal rate is not a good idea since simetimes there might be a slight difference between system clock, For example

  ***THIS IS BAD***
  #define SERIAL_PERD 100000  // Serial tx/rx check period, 100 ms is 10 Hz
  ---------
  self.comm_freq = float( rospy.get_param('~odom_freq', '10') ) # hz of communication
  

Version

1.0

Author

Ewing Kang - EwingKang

License

GPLv3

TODOs

[] github wiki