Chocolate Box – Infrastructure / Logic

As mentioned previously, the central part of the project is represented by the Infrastructure and Logic, both performed on a Raspberry Pi 3, powered by Raspbian.

The two pieces of software that I’m using are Node-RED for the automation and Mosquitto for the signaling.

Let’s see them in details:

Node-RED

Node-RED is a programming tool for wiring together hardware devices, APIs and online services in new and interesting ways.

It allows representing every piece of your automation system with a node, configuring its behavior, and connecting it with all the other elements of your system. It also allows to program functions, manipulate the messages, set delays, and so on.

There are a ton of libraries to operate with many different devices and services. For this project I used just a handful of them:

Let’s see how they are interconnected and configured:

Alexa-local node

The configuration of this block is trivial. It’s just enough to give a name to the node, and it will be the name that Alexa will list in the discovered devices. After this node is deployed it’s necessary to give an “Alexa, discover devices” command to make it appear in the smart home section of your Alexa as a new device.
You can find the Alexa-local node on GitHub or in the install palette on Node-RED.

Function “Open” node

This function takes the message coming from Alexa and manipulates the payload to control the opening of the box.

The function is extremely simple:

Delay node and Close Template node

The first node introduces a 20-second delay before triggering the close template.

The close template is defined as follows:

Once again, this is nothing fancy.

MQTT-out node

As you might have understood, the only two messages that this node receives are true, when it’s time to open the box, and false, 20 seconds later, when it’s time to close it.

These messages are sent to an MQTT broker with the topic box_open as follows:

And that’s all it’s needed on Node-RED to make it work. Let’s now see how to tell the Arduino board to open and close the box.

Mosquitto

To interface Arduino to Node-RED I went for the most appropriate protocol: MQTT. This protocol requires a broker, a piece of software that receives signals from the many different agents, and dispatches them to the appropriate recepients.

And this is all you need to know to make it work on our Raspberry Pi. You simply install Mosquitto and make sure it’s configured as your MQTT broker on Node-RED and on the Arduino sketch. Voila!

Downloads and resources

You can find the Node-Red Flow, and all the rest of the resources on this GitHub repository.

It’s all released under GPL!

That’s all for today, in the next posts I’ll show you how to configure Alexa, how to build the box, the circuit board, and configure Arduino.

Stay tuned, and feel free to contribute!

Chocolate Box – Overview

The Chocolate Box project is made of many different parts, which are summarised by this diagram:

Human Interaction

The interaction happens though Alexa, Amazon’s virtual assistant. In my case, any of the many Echo devices in my house can pick up the request and activate the whole flow.

Infrastructure / Logic

This part connects the signal coming from Alexa to the actuator device inside the box. The hardware for this part is a Raspberry Pi 3, and it’s powered by Node-Red for the automation system and Mosquitto as MQTT Broker.

Actuator

The actuator device is based on Arduino MKR1000, and two servos. All embedded into a cardboard box, disquised as a Christmas present.

In the upcoming posts of this series I’ll break down in details every block, and all the configs will be made available too.

Stay tuned!

Chocolate Box – intro

I built an automated chocolate box that sits nicely under my Christmas tree, and it’s operated by Alexa. It took me a few hours to figure out all the bits and pieces, but instead of working on a massive post with all the details, I’m breaking down this project in several pieces and several posts.

Stay tuned and enjoy the ride!

Oh, by the way, this is the final product:

I started this blog back in 2006 as a tech blog, so I feel this project a little bit as a return to the origins.

Turning a Bug into a Feature

I recently had a chat with my physiotherapist to fix the neck pain I have experienced more and more in the past few months.

He advised me to stretch my shoulders and neck. Better than a single 30-minute stretching session at the end of the day, he advised taking frequent 5-minute breaks during the day.

I recently discovered that the PIR sensors into Philips Hue devices require active movement, and not actual presence to be activated. This means that when I sit at my desk for an extended period of time, the sensor feels like nobody is in the room.

It was just enough to turn the timeout time for the sensor in my studio to 25 minutes, and I now have a perfect system that fixes both of m problems during the day: stretching my neck and shoulders and fully automate the lighting of my studio.

When I walk into the room, the lights turn on, and after 25 minutes after I leave the room, the lights turn off. When I’m working, if the lights turn off, it’s about time to stand up, stretch a little bit, drink a glass of water, and get back to work.

I also have a complicated node-red config that allows me to trigger on and off the sensor, so to avoid the awkward darkness during extended video meetings. But this is another story, so I’ll save it for another post.