This documentation is streamlined from an original project that uses Amazon Alexa to accept voice commands to move a robotic arm whose motors are controlled by an Arduino, to play a sliding puzzle. With application specific parts stripped away, this repo focus on a generally referrable and reusable structure/template that achieves: duplex communication between Arduino and Alexa Skill via HTTP API call.
More specifically, an Arduino Webclient will be set up to query states from or update new states to Alexa Skill running on Amazon Lambda function through a defined API in JSON format.
Here is a short video demo
Hardware:
- Arduino Uno ( Mega, Yun, etc. definitely works as well, but this doc is more specific to Arduino Uno whose network capability is limited, so if you unfortunately chose Arduino Uno like I did, read on :) )
- Arduino Ethernet Shield
- (optional) Amazon Alexa Echo
Environment:
- Node.js 6.10 LTS (recommend install Node Version Manager for nodejs version switching)
- Install and Setup ASK CLI
Amazon Account:
- Register for AWS account
- Register for Amazon Developer Account
- If you're able to assign your Arduino with a globally addresssable, static IP address; or if you are able to configure the router forwarding table under which your Arduino resides, then I'd strongly recommend using MQTT Subscribe&Publish Standard. Here is an MQTT client library for Arduino Ethernet Shield.
- In another word, this doc is ONLY for those who wants to duplex communication but whose Arduino can only connect to an institutional network (e.g. university, company) or private network that relies on a proxy server or NAT server for internet access.
- A quick way to find out, is to look up ip address info, click on
Get IP Detailand look atAssignmentfield to see whether it's statically assigned IP.
- A quick way to find out, is to look up ip address info, click on
- This doc assumes basic understanding of
- how a custom Alexa Skill works
- concepts like Interaction Model, Utterance, Intents, Slot Type
- If you are totally new to developing Alexa skill, it is strongly recommended to go through a few examples for conceptual familiarity.
- By the time of development, only
ask-sdkversion1 exits, but by the time of writing, there has been a newer release, still, no worry at all, migration usingask-sdk-v1adaptoris simple! Plus the update ofask-sdkonly affects minimal code in Alexa skill, the rest of the steps, configuration procedure remain unchanged.
-
STEP 0: System Architecture
-
STEP 1: Alexa Skill built with
ASK-SDKandASK CLI -
STEP 2: Implement Web Server in another Lambda Function
-
STEP 3: Configuring
Amazon API Gateway- API call handling hooked up with Lambda Functions
-
STEP 4: Invoking between Lambda Functions
- to modulate and decouple the web server logic from that of the Alexa skill's core logic
-
STEP 5: Configuring
Amazon CloudFront- accept HTTP request instead of HTTPS, while avoiding
cdnfor JSON request
- accept HTTP request instead of HTTPS, while avoiding
-
STEP 6: Implement Arduino WebClient
- to parse JSON response and initiate HTTP request
-
(Additional) A hacky way to keep Alexa Echo waiting for minutes without ending the session
This write-up doesn't put completeness on top, but instead it give disproportionate emphasis/details on things matter or things I wish I could have read or have as a guide during my development.
If you find this write-up useful and intend to wrap it into a more full-fledged documentation, please feel free to fork this repo or submit any issues. Any Pull Request are welcomed!
Plus, all code and markdown files are all under MIT license.
- Post migration source code (Alexa Skill re-implementation) using
ASK-SDK v2&& necessary documentation. - Noob friendly IAM Role Crash Course, especially highlighting necessary decisions/permissions associated with security concerns.