“Hello, Windows IoT Core” as Steve Texeira wrote few days ago !
The last stable release of the new IoT (embedded ?) OS for makers (and professionals ?) is out but of course I couldn’t wait the last days to start using it so I already played with it in a lot of demos and sessions code.
Who follows me knows that other than interacting with real world using sensors from Raspberry Pi 2, my great love is share these data and send them to the Cloud ! I already spoke and coded about using Pi 2 to send data to the awesome Event Hubs service (in the Microsoft Azure Service Bus family) and you can find my material on GitHub and SlideShare related to my “Telemetry with Windows 10 IoT : from Raspberry Pi2 to Event Hubs” session.
Of course, who follows me knows that I’m currently developing and updating (since about 2 years) the M2Mqtt library as part of the Eclipse Paho project (starting from October 2014) as the main MQTT client implementation for all .Net platforms (full, compact and micro) and WinRT (Windows 8.1 and Windows Phone 8.1) but … what about Windows 10 and Windows IoT Core ?
Don’t worry guys ! The last stable release (4.2.0.0) in its WinRT version works great on the new awesome Microsoft OS from the PC to the Raspberry Pi 2 thanks to the UWP (Universal Windows Platform) programming model. You can use the last Nuget package for sending telemetry data from your embedded system, your PC or smartphone writing only one application !
In the next few lines, I’ll describe how to use this package to send simple temperature value from my home. I’d like only to show how simple is to use the library and give you the right tools to create you IoT solution MQTT based for a Windows IoT Core system !
The hardware : Raspberry Pi2 and TMP102
Of course, it’s possible to develop an Internet of Things solution without an embedded system. Yes, of course ! I’m not crazy but I don’t want to discuss about this scenario (IoT is very huge you could think) during this article that will be focused on using an embedded hardware like the Raspberry Pi 2 with latest release of Windows IoT Core. Online, you can find a lot of article and official documentation on how to setup your board so I can consider this step already done for you !
Regarding the telemetry scenario … what is the easiest data to send … a temperature value of course ! 🙂
For this, I chose a temperature sensor I already had … the TMP102 from Texas Instruments in the related breakout board from Sparkfun. The connections to the Pi2 are so simple because the sensor supports I2C protocol and referring to the Pi 2 pinout we have (left Pi 2 pin and right TMP102 pin) :
- PIN 1 (3.3V PWR) –> V+ (power)
- PIN 3 (I2C1 SDA) –> SDA (I2C data)
- PIN 5 (I2C1 SCL) –> SCL (I2C clock)
- PIN 6 (GND) –> GND (ground)
- PIN 9 (GND) –> ADD0 (for I2C address selection)
You can leave the ALT pin of TMP102 not connected to Pi 2 because it’s for alert notifications from the sensor if temperature is higher or lower than a specific threshold. If you chose to use it, you need to connect it to a GPIO pin on the Pi 2 to receive an interrupt (Gpio Value Changed event) on alert.
Create UWP app and add the M2Mqtt package
Using Visual Studio 2015, it’s so simple to create a new UWP application from “File –> New Project –> Windows –> Universal” as showed in the following picture.
To develop on Raspberry Pi 2 and using all IoT features like Gpio pins, I2C and SPI devices, you need to add the Windows IoT Extensions for the UWP under references and selecting “Select Universal –> Windows –> Extensions”.
Now you are able to interact with all your Pi 2 pins and sensors you can connect to it but what about the M2Mqtt library for sending data to the Cloud using MQTT protocol. As I said, the library is available as Nuget package without need to recompile it from source code (it’s open source ! you can do it if you want !). Adding the package to your project is so simple under “References -> Manage NuGet Packages”, search for “M2Mqtt” and click on “Install”.
At the end of the installation you can find the package under the “Solution Explorer” window.
Time to code !
Now it’s time to code our application, reading from the temperature sensor and publish data to an MQTT broker. The first step is to interact with TMP102 sensor to get the temperature value. To do this I already wrote the driver you can find on GitHub here and include in your project.
To be much simpler as possible I wrote my code in the MainPage constructor to instantiate the TMP102 class and the MqttClient class from M2Mqtt library and to launch a task that reads the temperature and send it to the broker every 5 seconds. Of course, you have to write better code (for example using a Backgroud IoT Task instead of an UWP app) … but this sample is useful to show you how to integrate M2Mqtt in your project and how it’s simple to use it.
public sealed partial class MainPage : Page
{
private TMP102 tmp102;
private MqttClient client;
public MainPage()
{
this.InitializeComponent();
this.tmp102 = new TMP102();
this.client = new MqttClient("test.mosquitto.org");
this.client.Connect(Guid.NewGuid().ToString());
Task.Run(async () =>
{
bool isOpened = await this.tmp102.OpenAsync();
while (true)
{
float temperature = this.tmp102.Temperature();
string json = "{ temp : " + temperature + " }";
this.client.Publish("/pi2mqtt/temp", Encoding.UTF8.GetBytes(json));
await Task.Delay(5000);
}
});
}
}
To simplify the solution I decided to use the publically available Mosquitto broker at this address : test.mosquitto.org
The above example is very simple. First you have to create TMP102 and MqttClient instances (for the last one you need to specify the broker host name). The second step is to connect to the MQTT broker and start the task for reading and sending data. Inside the task, every 5 seconds, the temperature value is read from the TMP102 instance and it’s published to the broker on the “/pi2mqtt/temp” topic in JSON format (as payload of the MQTT message).
Now that your project is sending value to the broker, the simplest way to see you are publishing data is to use another MQTT client, for example the mosquitto client and subscribe to the same topic to receive data. After installing mosquitto, you can start the mosquitto subscriber client in the following way :
mosquitto_sub -h test.mosquitto.org -t /pi2mqtt/temp -q 0 -d
Now … you should see data received from you Raspberry Pi 2 !
Conclusion
Of course, it isn’t an article to deep into Windows IoT Core, MQTT protocol, M2Mqtt library and so on. I’d like only to show how you have an “out of box” way (as the M2Mqtt library) to integrate your embedded system into a huge MQTT solution and you can develop it using Windows IoT Core as OS or in general your Windows 10 system !
DEVEXPERIENCE 