Skip to content
/ blinky Public

A web-based touchscreen face for HRI research.

License

Notifications You must be signed in to change notification settings

hcrlab/blinky

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

31 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Blinky

A web-based touchscreen face for HRI research. It has a ROS interface for changing the face and calling behaviors. Blinky is primarily designed and tested for Chrome on a Nexus 7 (2013) tablet.

blinky

Getting started

  • Install the latest version of node using nvm
  • Clone this repository to your catkin workspace and build it.
  • From the frontend folder, run npm install -g bower polymer-cli and then npm install && bower install
  • Run roslaunch rosbridge_server rosbridge_websocket.launch
  • Run polymer serve and go to the generated URL to see the face.
  • It's recommended that you add the app to the device's homescreen, so that the app can be run full-screen. In Chrome, tap the menu button in the top right (the "three dots") and tap "Add to Home screen." On Android 5.0 (Lollipop) and above, you can additionally pin the app using Screen Pinning to hide the notification bar.

You should see a blank face. To send a message to it, send a goal to its actionlib server, /blinky, of type blinky/FaceAction:

rostopic pub -1 blinky/goal blinky/FaceActionGoal "header:
  seq: 0
  stamp:
    secs: 0
    nsecs: 0
  frame_id: ''
goal_id:
  stamp:
    secs: 0
    nsecs: 0
  id: ''
goal:
  display_type: 'displayMessage'
  h1_text: 'Hello world!'
  h2_text: 'How can I help you?'"

Deploying

These instructions describe how to deploy Blinky to an Apache server running on the robot computer. We assume that a Nexus 7 tablet is connected to the laptop via USB.

Build the website

cd frontend
polymer build
sudo cp -r build/default /var/www/blinky

Set up Apache

Install Apache if you don't already have it: sudo apt-get install apache2

In our setup, we are going to serve the website from port 8000. The URL will be something like robot.university.edu:8000/ This allows us to serve the website from the robot's computer via USB, thanks to Chrome's port forwarding / mobile device debugging feature.

See DigitalOcean's guide to configuring Apache for a great primer on configuring Apache.

Create the file /etc/apache2/sites-available/blinky.conf:

Listen 8000
<VirtualHost *:8000>
  ServerAdmin webmaster@localhost
  ServerName robot.university.edu
  ServerAlias robot
  ServerAlias localhost

  DocumentRoot /var/www
  <Directory />
    Options FollowSymLinks
    AllowOverride None
    Order Deny,Allow
    Deny from All
  </Directory>

  <Directory /var/www/blinky>
    Options Indexes FollowSymLinks MultiViews
    AllowOverride None
    Order allow,deny
    allow from all
    require all granted
  </Directory>

  ErrorLog ${APACHE_LOG_DIR}/error.log

  # Possible values include: debug, info, notice, warn, error, crit,
  # alert, emerg.
  LogLevel warn

  CustomLog ${APACHE_LOG_DIR}/access.log combined
</VirtualHost>

Enable the blinky site using sudo a2ensite blinky.conf. Then, restart the Apache service with sudo service apache2 reload.

Load the page on the tablet

You should be able to access the webpage from the regular internet by going to robot.university.edu:8080/blinky. However, we want the face to be robust to loss of wifi. Instead, we load the page directly from the robot's computer over USB.

To do this, open an SSH connection to the robot's computer using the -X option and run google-chrome

ssh -X robot
google-chrome

On the tablet, enable USB debugging (see Remote Debugging Devices). From Chrome, go to chrome://inspect. You should see the Nexus 7 in the device list. Go to port forwarding and forward port 8000 to localhost:8000 and port 9090 to localhost:9090, which is necessary for the rosbridge websocket connection (see Remote Access to Your Local Site).

Behaviors

displayMessage(h1_text, h2_text)

This will make the robot display a message to communicate to users. The robot can display big text (for legibility at a distance), or just regular text (to fit more content on screen), or a combination of both.

askMultipleChoice(question, choices)

This will make the robot ask a question on screen. choices is a list of strings, each of which will be displayed as a button. The choice the user selects will be returned as the actionlib result. If there is only one choice, the button will be colored, otherwise they are grey.

If you are interested in helping us create more behaviors, please let us know!