-
This package provides a set of ROS packages for using a Parallax Arlo Platform robot to run all of the demonstration projects for the Robot Operating System (ROS) based TurtleBot
-
This package also installs a node.js based front end GUI for the robot, complete with Twilio and Pushover integration and push button access to ROS functions:
- The GUI also includes a remote control "Teleop" interface:
Here is a demonstration video of this code on my ArloBot: http://youtu.be/7qJaA6K_WPE
First you need to build a robot!
The Parts List is now a GitHub Wiki page.
Check out my blog: ArloBot Build Index
And also jump on the Parallax Forums and start searching and asking questions!
There is no one complete source on how to do this, but we will all help you out. Ask questions in the Parallax Forums, on GitHub, and on my blog. We will write more documentation as we answer questions, and I hope you will also write instructions when you do your build!
Once your robot is built, you can use this package.
Arlobot operates on ROS Indigo which requires Ubuntu 14.04 LTS, or Xbuntu or Lubuntu of the same version. I personally recommend Lubuntu.
If you put a fresh copy of Lubuntu on your robot's laptop then you can use the quick install script below.
There is now a script to install everything. Just run:
bash <(wget -qO- --no-cache https://raw.githubusercontent.com/chrisl8/ArloBot/master/setup.sh)
and you are done! (Remember, you must use Ubuntu 14.04 LTS for ROS Indigo! This package does not work on Jade yet!)
Be sure to read the instructions that the script will print at the end about editing the config files in ~/.arlobot/
To update your code just run the same script again and it will pull down and compile anything new without erasing custom settings.
Please note that you will need the code to run on your Propeller board. This is stored in the "Propeller C Code for ArloBot" folder. Details on the Propeller code and setup are here: http://ekpyroticfrood.net/?p=165
Run:
~/catkin_ws/src/Metatron/startRobot.sh
and point your web browser at the URL it gives you.
If you use Ubuntu or Lubuntu there should also be a desktop icon on the robot's desktop that you can run to do the same thing and bring up this web page on the robot itself.
If you have a desktop or another laptop computer running Ubuntu that you just want to run RVIZ, rqt_graph, etc. on, you can run this script to set up enough of ROS to do that, without attempting to compile the robot code.
The "workstation" install uses the latest "Jade" version of ROS, so it will work on Ubuntu 15. Just run:
bash <(wget -qO- --no-cache https://raw.githubusercontent.com/chrisl8/ArloBot/master/workstation.sh)
and your system will be set up to use as a "remote" station. To update your code just run the same script again and it will pull down and compile anything new without erasing custom settings.
Complete setup and usage instructions are at my blog: http://ekpyroticfrood.net/?p=162
It is a good idea to irst test your serial interface to the propeller board:
cd ~/catkin_ws/src/Metatron/scripts/
./direct2PropSerialTest.sh
This will make a direct serial connection to the Propeller Activity board.
It will reset the Prop board and then start spitting out:
i 0
Then paste the line suggested by the script to initialize the program and it should start sending odometry info in the form of:
o 0.000 0.000 0.000 0.000 0.068 0.000 {"p0":135,"p1":90,"p2":78,"p3":78,"p4":107,"p5":34,"p6":15,"p7":11,"p8":16,"p9":67,"p10":77,"p11":120,"p12":9,"p13":10,"i0":1991,"i1":212,"i2":153,"i3":82,"i4":99,"i5":24,"i6":25,"i7":12}
With the occasional:
s 1 1 0 100 10 12 0.05 0.06 0 0
You may have fewer "p#" and "i#" instances. Those are the distance readings from your PING an IR sensors. They should change as you move around your robot. The "s" line is handy because it tells you about some of the robot's decisions.
1 = True and 0 = False
The meanings of each numbers are:
- Safe to Proceed
- Safe to Recede
- Escaping
- Forward speed limit
- Reverse speed limit
- Distance sensor with lowest number
- Left motor voltage
- Right motor voltage
- Cliff detected
- Floor obstacle detected
If you want to get really fancy you can even send it twist commands from the terminal too! Just remember to turn on the motors first for that to work!
Twist command format:
s,0.0,0.0
where the first number is the linear meters per second and the second number is the angular radians per second of a standard ROS "Twist" message.
Slow Forward:
s,0.100,0.000
Slow Reverse:
s,-0.100,0.000
Slow turn left or right:
s,0.00,0.50
s,0.00,-0.50
Depending on what you want to do there are different ways to "bring up" the robot with just ROS.
These are the "recipes" that are well tested so far:
Basic TeleOp with 3D sensor use:
roslaunch arlobot_bringup minimal.launch --screen
<New Terminal>
roslaunch arlobot_teleop keyboard_teleop.launch
<New Terminal>
roslaunch turtlebot_bringup 3dsensor.launch
<GUI based Terminal>
roslaunch arlobot_rviz_launchers view_robot.launch
Tests from this setup:
Set your Global Options->Fixed Frame to "odom
Drive and see if the robot appears to move properly on the grid.
Turn on LaserScan and set the Decay Time to 650
Move around, spin in circles and see if you get a reasonable picture of the room.
Turn off Laser Scan and turn on Registered DepthCloud to see if you get a picture of the room overlaied properly onto the 3D virtual world in RVIZ.
Remote Control with an xBox 360 joystick:
http://ekpyroticfrood.net/?p=115
roslaunch arlobot_bringup minimal.launch --screen
<New Terminal>
rosparam set /joystick/dev "/dev/input/js0"
roslaunch turtlebot_teleop xbox360_teleop.launch --screen
Follow a person or object:
http://wiki.ros.org/turtlebot_follower/Tutorials/Demo
roslaunch arlobot_bringup minimal.launch --screen
<New Terminal>
roslaunch arlobot_bringup follower.launch --screen
Gmapping Demo (SLAM Map building):
http://wiki.ros.org/turtlebot_navigation/Tutorials/Build%20a%20map%20with%20SLAM
roslaunch arlobot_bringup minimal.launch --screen
<New Terminal>
roslaunch arlobot_navigation gmapping_demo.launch --screen
<New Terminal>
roslaunch arlobot_teleop keyboard_teleop.launch
<GUI based Terminal>
roslaunch arlobot_rviz_launchers view_navigation.launch
When you are done, save your map!
rosrun map_server map_saver -f ~/rosmaps/my_map1
Tests from thsi setup:
Make sure that obstacles in the Asus view are shown in the local costmap
It is possible to map walls, while the 3D is ignored by the costmap!
I find this is caused by the max_obstacle_height being set below the 3D Sensor's height
in costmap_common_params.yaml on the "scan:" line
AMCL (Navigating the map we built above:
http://wiki.ros.org/turtlebot_navigation/Tutorials/Autonomously%20navigate%20in%20a%20known%20map
roslaunch arlobot_bringup minimal.launch --screen
<New Terminal>
roslaunch arlobot_navigation amcl_demo.launch map_file:=~/rosmaps/my_map1.yaml
<GUI based Terminal>
roslaunch arlobot_rviz_launchers view_navigation.launch --screen
NOTE: This is still in progress. It works pretty well, but needs a little tweaking.
Autonomous Map Making (Exploration):
roslaunch arlobot_bringup minimal.launch --screen
<New Terminal>
roslaunch arlobot_explore gmapping_explore.launch --screen
<New Terminal>
roslaunch arlobot_explore exploration_planner.launch --screen
<GUI based Terminal>
roslaunch arlobot_rviz_launchers view_navigation.launch
<New Terminal>
roslaunch arlobot_explore arlobot_explore.launch --screen
When you are done, save your map!
rosrun map_server map_saver -f ~/rosmaps/my_map1
Please report an issue for any problems or if you need me to clarify anything!