Vocal: Voice Automated Lights

Vocal is a simple application written in Python that simulates the lights of a simple house with three rooms, following the client-server model. Each room is represented by a LED on a BeagleBone Black, that is turned on or off after a voice command given by the user.

Setting up the BeagleBone Black

First you must configure the BeagleBone Black hardware. For Vocal, you'll need three different colored LEDs, three resistors and a few jumpers. We've implemented like this:

To proceed, you'll need to install the Adafruit Beaglebone I/O Python API module on your BeagleBone.

Also, in the file board_assets.py, you'll find a dictionary named house. It has as keys all the rooms and as values the GPIO pin number associated to that room. You can change this configuration to the one that suits more your needs, but remember to put LEDs, jumpers and resistors connecting everything properly.

Setting up the server

For running the server, transfer to your BeagleBone the files vocal_server.py and board_assets.py.

Notice that in vocal_server.py, right at the beginning of the file theres two constants: HOST and PORT. You must type there the IP address of your BeagleBone to the HOST (e.g. "192.168.0.24") and a valid port (recommended a number up to 1024) number to PORT (e.g. 6000).

After this, simply run: python vocal_server.py and you should see something like this:

The server configured the GPIO and is ready to receive socket connections and perform the voice commands.

Setting up the client

The client will stay on your machine. To set it up, you'll need to install PyAudio and SpeechRecognition, as described at requirements.txt. You can do this simply by:

• pip install -r requirements.txt

Before running, you'll have to provide host and port for the socket in vocal_client.py, the same way you've done on vocal_server.py.

After this, and having successfully installed all the requirements, you can execute the code typing on your terminal: python vocal_client.py.

If everything goes well, you should see a screen similar to this:

Its okay if you see those messages that begin with "ALSA" on Ubuntu. If you want ro remove them, check this out.

Running Vocal

From the screen above showed on "Setting up client" section, Vocal is already running. Now you can hit ENTER and the program will ask for your voice input.

For this example I said "all on". Vocal receives this input, treats using the SpeechRecognition and sends the instruction via socket to the server. After the server performs his actions, it sends back a message informing wheter the command worked or not.

Here's how it looks like from the server perspective after the client sends the instruction:

The server accepts the connection and sends the "house" status to the client (more on this later). After he decodes the instruction received, he performs the necessary actions.

This is how it looked like on the BeagleBone after the server executed the voice command:

Since the command was "all on", all the rooms (LEDs) were turned on.

The log messages

Periodically (in this example, every 1 minute) the server will send to the client a message with the status of the "house" (if lights of all rooms are turned on or off). The client will receive this message and will store them in a log file.

Here's the output of this log file after some minutes of execution:

Enjoy yourlsef!

Feel free to test, fork and contribute to this project! If you have any doubts, contact us with:

• Lucas Gomes Dantas - [email protected]
• Lucas Aléssio Anunciado Silva - [email protected]

Authors

Lucas Dantas	Lucas Aléssio