gypsybud / avs-alexa-iot-intel-speech-enabling-kit

Learn how to build voice enabled IoT solutions with the Intel Speech Enabling Kit & AWS

What Is It?

With your Alexa/ Voice Service (AVS) enabled device, you can create Alexa skills that let's you interact with hardware devices, such as reading sensor data or activating a relay, all with the command of your voice!

How It Works

An Alexa skill converts spoken commands into computing functions, which are defined as an Amazon Web Services (AWS)* Lambda. The computing functions can be used to communicate with your IoT device using AWS IoT features.

Components

Hardware

1. AVS Device (Rasberry Pi* + Intel® Speech Enabling Developer Kit)
2. Sensors/Actuators (Arduino 101* (branded Genuino 101* outside the U.S.) + Grove* Starter Kit)

Software

1. Alexa Voice Service (brings voice capabilities to your device)
2. Alexa Skills Kit (defines the voice interaction model e.g. intents & utterances)
3. AWS Lambda (compute functions that execute on defined voice commands)
4. AWS* IoT (provides channel to communicate with your IoT device)
5. MRAA/UPM (hardware abstraction library for reading/writing to sensors/actuators)

Steps

1. Setup your Intel® Speech Enabling Developer Kit

• Follow this tutorial to setup your Intel® Speech Enabling Developer Kit with AVS.

2. Connect sensors/actuators and setup MRAA & UPM

1. Plug in an Arduino 101 to your main computer and upload the StandardFirmata sketch.
2. Attach a Grove Shield to the Arduino 101:

• Attach a temperature sensor to the A0 port.
• Attach an LED to the D4 port (can also be a relay).

3. Plug the Arduino 101 into your Raspberry Pi running AVS.
4. Install MRAA & UPM:
   • npm install -g mraa
   • npm install -g jsupm_grove
   • Note: Make sure you have all dependencies are resolved, such as node-gyp, nodejs-dev, build-essential. etc.
5. Identify the serial port where the Arduino 101 is connected (usually /dev/ttyACM0).
6. Try some MRAA/UPM code samples to make sure you are able to interact with the Arduino 101.

3. Connect your device to AWS* IoT

1. We want to send temperature data to AWS IoT so our Lambda function can access it. First, we need to create a new AWS IoT 'Thing'. Follow this tutorial for details.
2. With the newly created 'Thing', copy over the certificates and keys to your Raspberry Pi.
3. Run '/iot/index.js' and check on the AWS IoT dashboard to ensure the data is being updated.

4. Create an Alexa* Skill

1. Go to to create a new Alexa Skill.
2. Fill in the Interaction Model details with the data under /skill/ e.g. IntentSchema.json, SampleUtterances, Slots.
3. Save your progress and resume later. You can complete the rest of the configuration after you've created a Lambda endpoint.

5. Connect your Alexa* Skill to an AWS* Lambda

1. Create a new Lambda function
2. Select 'Alexa Skills Kit' as a trigger.
3. Upload the Lambda function code:
   1. Under /lambda run npm install to install the node_modules.
   2. Zip all the files by running npm run zip
   3. Upload the output file index.zip to your Lambda and save.
4. Test your Lambda with the Alexa Skill with the Service Simulator e.g. Enter "What's the Temperature" as an example utterance.

6. Try out your skill!

1. Start the data collection script node /iot/index.js
2. Start your skill by saying 'Alexa, open NAME_OF_SKILL'
3. Ask it "What's the temperature?" to get the temperature sensor data.
4. Say "Turn On" to turn on the connected LED.
5. Enjoy!

Troubleshooting Tips

• Ensure your security policies have the right permissions.
• Enable AWSIoTFullAccess for your Lambda Execution Role.