Introduction: Controlling Devices Over the Internet of Things - Part 1

By

The Internet of Things (IoT) is one of the buzz words nowadays and as a software developer I’m curious about it and will try to put in here what I believe is the easiest way to achieve it. This means, of course, that you may know different ways much more simpler and effective than what I know.

First of all, I have to admit that I’m not a hardware person, in fact I´m really bad at it. Fortunately for me and others in the same situation, there are plenty of tools you can use to build your electronics. In my case, I use:

  • Arduino Breakout board.
  • Intel Edison.
  • Seeed Studio Grove shield.
  • Grove LED.
  • Grove Button.

Just as a quick reference, Intel Edison in conjunction with Arduino Breakout board provides the same software and hardware interfaces as the Arduino UNO but with more power as it is running Yocto, a linux-based distribution created for embedded devices. On the other hand, Grove is a tool set created by Seeed Studio to give makers and inventors the ability to create devices with simple building blocks in a plug-and-play fashion. You can say it’s like Lego for electronics.

On the software side, I used Node-Red, which is a tool that helps you orchestrate flows by visually wiring your processes. A web browser will be used to test the service and manipulate the LED.

In this first part, I’ll show you how to connect the circuit and send commands to turn on or off a LED connected to the Grove shield through an HTTP request. This, of course, will allow you to interact with the LED when you’re in the same WiFi as the Edison. On a second part, I’ll show you how to connect the circuit to a cloud IoT instance and send commands to the Edison via Internet.

Step 1: Before You Start

Make sure the following is already installed.

  1. Edison is configured to connect to WiFi. If you haven’t setup your Edison, follow the steps in Intel Edison site.
  2. You can login to Edison. Once the WiFi is configured, you can login using SSH. I found this tutorial very useful.
  3. Node-Red is installed in Edison. Follow instructions in the Node-Red documentation.
  4. Node-red-contrib-gpio module is installed in Node-Red. Follow instructions in the official documentation in GitHub.
  5. Ionic is installed on your computer. Follow instructions in Ionic’s web site.

Step 2: Assembling the Circuit

The assembly is quite simple:

  1. Plug the Edison and the Grove shield in the Arduino Breakout.
  2. Plug the cable to the Grove LED board.
  3. Plug the cable to the Grove shield, I chose slot D6 but you can use any digital pin.
  4. Plug the micro USB cable to the Arduino power and to the computer to turn it on.

Step 3: Creating the Flows in Node-Red

Here, I’ll assume you have Node-Red up and running in your Edison and available through your web browser.

In Node-Red, you will create the flow that will handle a GET HTTP request, extract the parameter of the action on the LED, turn on or off the LED, format a message and send it back to the client in the form of a HTTP response.

The complete flow should like like this one. Follow these steps to complete it.

Step 4: HTTP Request

In the Input drawer in Node-Red select the HTTP node, drag and drop it onto the canvas. Double click on the node in the canvas and specify the following parameters:

  • Method: GET
  • URL: /led/:state
  • Name: ActionOnLED

Don’t miss the colon (:) in the URL. After setting these attributes, click on Done.

I used the GET method because at first I wanted to try using a web browser. Since this is a command, I think a POST method suits better for the purpose but I’ll leave the correct architecture terms on your end.

Step 5: Extract the Action From the HTTP Request

As you can tell in the URL above, the action is passed as the last portion of the URL as a path parameter. Because of this, we need to extract it. Node-Red gives you the chance to get variables from the request using the standard Express API.

From the Function drawer select the function node and drag and drop it into the canvas. Double-click on it and put the following JavaScript code.

if (msg.req.path == "/led/on") {
    msg.payload = 1;
    msg.httpMsg = "On"
} else if (msg.req.path == "/led/off") {
    msg.payload = 0;
    msg.httpMsg = "Off"
} else {
    msg = null;
}

return msg;

After entering the code, click on Done.

Note that there’s a new property in the message to pass a readable string which will be used to notify the client.

Now that the function is complete, wire the two nodes each other. On the HTTP request node, click on the small gray square and then drag and drop it on the gray square on the function. The flow is now wired, just this simple!

Step 6: Control the LED

It’s time to add a node to control the LED on the shield. In the Input drawer select the gpio node and drag and drop it on the canvas. Double-click on it to specify the attributes for this node.

First, you will need to specify the board type you have. For this, click on the pencil icon in the Board field. In the resulting wizard, specify your board for the Nodebot. In this case select Galileo/Edison and set a name. Click on Update to save the changes.

Once the board is configured, specify the attributes to control the LED:

  • Board: The board you just configured.
  • Type: Digital(0/1).
  • Pin: As I mentioned above, I used pin 6.

After setting these attributes, click on Done and wire this node to the function node created above.

Note the after clicking on Deploy, Node-Red will try to connect the physical board. When it’s successful it displays the connected!!! message below the node.

Step 7: Format the Message for the Client

From the Function drawer select the function node and drag and drop it into the canvas. Double-click on it and put the following JavaScript code.

msg.payload = {
    message: "LED was turned " + msg.httpMsg,
    led_state: msg.payload
}
return msg;

After entering the code, click on Done and wire this node to the other function node created above.

Step 8: Sending Back a Response

Since this request is initiated with an HTTP request, you must finish it with an HTTP response. For this select a HTTP Response node in the Output drawer and drag and drop it on the canvas. Wire this node to the function that was created previously to form the message for the client.

Click Deploy to bring your changes to life.

Step 9: Testing Your Flow

Now that the flow is created, you can use your web browser, or any other program that handles HTTP requests like cURL or WGET, to test it.

Open your web browser and use either of these URL’s. Make sure your IP address is the one configured in your Edison. In my case, the IP address is 192.168.1.71.

Step 10: Limitations

I could only found one limitation on actual functionality here. In the scenario where you turn on the LED using the HTTP request directly from the web browser, the UI in the web application won’t reflect that state because the connection is asynchronous and disconnected. That is the web application is not notified of changes in the state of the LED.

To resolve this problem you could use different approaches like using WebSockets instead of HTTP requests to establish a bi-directional communication and have synchrony between the HTTP requests and the UI. Or maybe you could use MQTT as the protocol to subscribe and publish events.

In any case you will need to modify the flow in Node-Red to accommodate the new transport protocol and of course your JavaScript code.

IoT Builders Contest

Participated in the
IoT Builders Contest