hueper

This software is part of a study project done by fairylands and kryops at DHBW Stuttgart in 2013/14.

It provides a Node.JS + MongoDB backend and a real-time cross-platform mobile web frontend for extended control and automation of the Philips Hue lights, intended for usage on a Raspberry Pi together with an Arduino. Its features include:

Most of the existing functions of the Hue Bridge (control of lights, groups, user administration, firmware update)
Save Favorite color settings
Scenes
Flexible scheduling and automation framework
Environment sensors provided by the Arduino
Control through RFID/NFC tags
Network device recognition
Speech recognition (online with Google Speech API or offline with Julius)
Party mode for time- and sound-controlled light changes
All functions controllable via Socket.IO or a REST API, real-time synchronization between clients

Licensed under the Apache License. The complete documentation is available in the docs folder (German only!).

Installation

Complete hardware list in docs/planung/hardware.md (German)

Dependencies

Linux

sudo apt-get install git python build-essential nodejs mongodb arduino sox nmap julius
sudo usermod -aG dialout $USER

Windows

NodeJS: http://nodejs.org/download/
MongoDB: http://www.mongodb.org/downloads (start with C:\<Path>\bin\mongod.exe --dbpath "C:\<Data directory>")
Git: http://msysgit.github.io/
Phyton (Version 2.x!): http://python.org/download/
Arbitrary version of Visual Studio (or Express), e.g. http://www.visualstudio.com/downloads/download-visual-studio-vs#d-express-windows-desktop
SoX: http://sourceforge.net/projects/sox/files/sox/14.4.1/ (Install in C:\Program Files (x86)\sox-14-4-1 or add to PATH)
Julius: http://sourceforge.jp/projects/julius/downloads/60273/julius-4.3.1-win32bin.zip (Install in C:\julius-4.3.1-win32bin or add to PATH)

Project setup

(in Windows you may have to switch between the Git bash and the NodeJS command line!)

git clone https://github.com/SBejga/hueper.git
cd hueper/nodejs
npm install

Download acoustic model: http://www.repository.voxforge1.org/downloads/Nightly_Builds/AcousticModel-2014-05-15/HTK_AcousticModel-2014-05-15_16kHz_16bit_MFCC_O_D.zip and extract to julius/acoustic_model_files
Start NodeJS: node <Project path>/nodejs/server.js
Frontend is accessible at http://localhost:8080

Get Google API Key for speech recognition

In order to use the Google speech recognition engine you have to generate a key with a Google account. Please follow the steps provided here: http://www.chromium.org/developers/how-tos/api-keys

In the Google Developers Console / APIs activate Speech API
In Credentials / Public API access create a Server Key
Save it in the application settings

Arduino setup

Hardware needed:

Arduino UNO
Adafruit PN532 RFID/NFC Shield + Stackable Header Kit
Seeedstudio Grove Base Shield
Seeedstudio Grove Light Sensor (A0), PIR Motion Sensor (D8) and Sound Sensor (A2)

Installation:

Arduino IDE: http://arduino.cc/en/Main/Software
Select the right Serial Port
Include the library for the Adafruit PN532 RFID/NFC Shield: https://github.com/adafruit/Adafruit_NFCShield_I2C
Flash the arduino/arduino.ino sketch

Deploying on a Raspberry Pi

Hardware needed: External USB sound card with microphone (in /etc/modprobe.d/alsa-base.conf set options snd-usb-audio index=0)

Flash your Raspberry Pi with Raspbian
Connect it to the internet, make sure all packages are up to date
Copy raspberry/setup.sh to /home/pi
Make it executable and execute it (as root)
wait :)
reboot

Known Issues

After restarting the service on the Raspberry Pi the Julius speech recognition doesn't work any more
Android Browser sometimes fails to establish a Socket.IO connection. This may possibly be solved by upgrading to Socket.IO 1.0