Turn on an LED With Watson Conversation

What you'll need:

Zagros Artificial Intelligence Raspberry Pi Kit

Step 1:

You will need to have node already installed on your Pi. You may find NPM's rpio-gpio site helpful for syntax.

Step 2:

Run the command npm install rpi-gpio in terminal.

Step 3:

We started with the conversation.js file from Watson's TJBot example, and added the following lines before we instantiated our bot.

var gpio = require('rpi-gpio');

var pin = 7;

gpio.setup(pin, gpio.DIR_OUT);

The first parameter for setup() is the channel. Make sure to reference the RPi pin number and not the GPIO. The second parameter is the direction, DIR_OUT writes to pin #7. You can also change the name of your bot to something different. We picked "Bob" as it was less likely to be confused with other words.

// instantiate our TJBot!

var tj = new TJBot(hardware, tjConfig, credentials);

tj.configuration.robot.name ="Bob";

Step 4:

After the utterances part of the code add the following code for speech recognition.

var containsOn = msg.indexOf("on") >= 0;

var containsOff = msg.indexOf("off") >= 0;

var containsLight = msg.indexOf("light") >= 0;

//turns on light

if (containsLight && containsOn ) {

console.log("Turn on Light")

gpio.write(pin,true);

};

// turns off light

if (containsLight && containsOff ) {

console.log("Turn off Light")

gpio.write(pin,false);

};

Step 5:

Setup for the pins.

Step 6:

The complete node js code.