This project demonstrates the use of MQTT protocol for collecting environment data (pressure, temperature and humidity) with MQTT clients distributed on multiple locations, running on Raspberry Pi with Pi Sense HAT attached. libmosquitto is selected for implementing the MQTT publishers and subscribers in C/C++, while Eclipse Mosquitto is MQTT broker of choice for testing the code. In one of the examples, Home Assistant is the MQTT subscriber on a topic published by MQTT client written with libmosquitto.
Four MQTT clients are included in the project:
- mqtt_sub: a subscriber to ambient data topic;
- mqtt_pub: a publisher of dummy/test ambient data (used for testing the MQTT setup);
- mqtt_pub_sense_hat: same as mqtt_pub, but it uses the sensors on Raspberry Pi Sense HAT for providing actual ambient data;
- mqtt_pub_ha_sub: the readings from Raspberry Pi Sense HAT are published in JSON format via Eclipse Mosquitto Broker. Home Assistant plays the role of the subscriber. A detailed description of this use case is presented here.
-
Client mqtt_sub. It subscribes to every topic on the broker that starts with home and ends with ambient_data ("home/+/ambient_data"). The loop for receiving MQTT messages runs in a separate thread. The received payload will end up in a FIFO queue. Additional worker thread will process every payload entry from the queue by calling the process_message() function. In this example, the function only prints the payload on a standard console.
-
The publisher mqtt_pub writes on the topic either dummy or real environment data it collects for its location. The client publishes the MQTT message in a loop.
Supported command line arguments:
-b <hostname/IP of the broker> default value: localhost;
-p <port number> default value: 1883;
-l <location> default value: location_<pid of the process>, ignored by mqtt\_sub if given.
The client will use the default values for the missing arguments.
The MQTT message carries control and payload data. The payload consist of: location name, temperature, pressure and humidity.
All MQTT messages are send with QoS (quality of service) flag set to 0, and retain field set to false. The clients neither support MQTT authentication nor they can establish a secure connection with the broker over SSL channel.
- Libmosquitto library;
- Libsetila for Pi Sense HAT envirement sensors support;
- CMake for building the project from source;
- Mosquitto broker for testing the clients functionallity.
The code should work on all platforms capable of running Linux with multithreading, and are supported by libmosquitto. Only the mqtt_pub_sense_hat is written for Raspberry Pi with Pi Sense Hat, althoug if will work on other embedded systems with LPS25H and HTS221 sensors attached.
Before building the source code on Ubuntu or Raspbian, make sure that libmosquitto-dev package is already installed. The package is available in the official Ubuntu and Raspbian repositories.
For Pi Sense HAT support, first install libsetila library. Check libsetilla installation manual for the instructions.
CMake tool is selected for building the clients from the source. It is an out-of-source build which requires a separate folder for the compiling process. Navigate to the project folder, create a new subfolder and jump inside:
#mkdir build && cd build
Execute the following command in order to create a building instructions for the release version of the binaries:
#cmake -DCMAKE_BUILD_TYPE=Release ..
or
#cmake -DCMAKE_BUILD_TYPE=Release ..
for debug version.
The last two commands will generate rules for building mqtt_sub and mqtt_pub binaries only. For getting mqtt_pub_sense_hat build, use additional argument -DWITH_PI_SENSE_HAT=ON:
#cmake -DCMAKE_BUILD_TYPE=Release -DWITH_PI_SENSE_HAT=ON ..
To include the Home Assistant example in the build, add -DWITH_HA_EXAMPLE as an argument of the cmake command.
After Cmake generated the build scripts, compile the clients with:
#make
In order to install the binaries in the bin folder under the project tree use:
#make install
There are different deployment scenarios depending where the MQTT broker and client will be hosted/running. One of the possible implementations is presented on the diagram below.
In this example, the broker and *mqtt_sub *run on a same Ubuntu 18.04 host or the collector node, while mqtt_pub_sense_hat clients will be started on two Raspberry Pi with Pi Sense HAT running Raspbian OS. Those are the publishing nodes.
Bild the project on the collector node to get mqtt_sub client binary. Next step is Mosquitto broker installation.
Mosquitto broker is available in the official Ubuntu repositories. Use the command below for installing the broker package:
#sudo apt-get install mosquitto
Configure Mosquitto. The default configuration is enough for using the broker with the clients.
Check if the broker is up and running. Seach for mosquitto string in the output of netstat command:
#sudo netstat -tlnp | grep mosquitto
The last command should produce the foloowing output:
tcp 0 0 0.0.0.0:1883 0.0.0.0:* LISTEN 9847/mosquitto
tcp6 0 0 :::1883 :::* LISTEN 9847/mosquitto
Start Mosquitto broker if it is not already running.
#sudo systemctrl start mosquitto.service
Optional, configure Mosquitto to start at boot automatically:
#sudo systemctrl enable mosquitto.service
Install the prerequisites required for building mqtt_pub_sense_hat on Raspberry Pi with Sense HAT. Build the binary from source.
MQTT broker should be already running on the system, waiting to manage MQTT messages. Start the subscriber client on the collector node. To do so, in the command line, navigate to the bin folder under the project folder and start the binary:
#cd <project_folder>/bin
#./mqtt_sub
Because it runs on the same host as the broker, no arguments for broker hostname and port are required.
After the subscriber is running, start the mqtt_pub or mqrr_pub_sense_hat processes on the Pi client(s):
#cd <project_folder>/bin
#./mqtt_pub_sense_hat -b mqtt-broker -l kitchen
Or with IP address of the broker machine (10.10.10.2 on the deployment diagram) as a value for the -b argument:
#./mqtt_pub_sense_hat -b 10.10.10.2 -l kitchen
On the second Pi, mqtt_pub will be started without providing the location info:
#./mqtt_pub -b mqtt-broker
Mqtt_sub will print the received ambient data in the console where is was started:
03.May.2020 14:55:17 [kitchen] t = 32.21[°C], p = 1001.39[hPa], H = 35.20[%rH]
03.May.2020 14:55:17 [location_13321] t = 25.30[°C], p = 995.30[hPa], H = 33.00[%rH]
03.May.2020 14:55:18 [kitchen] t = 32.20[°C], p = 1001.38[hPa], H = 35.20[%rH]
03.May.2020 14:55:18 [location_13321] t = 25.30[°C], p = 995.30[hPa], H = 33.00[%rH]
03.May.2020 14:55:19 [kitchen] t = 32.21[°C], p = 1001.37[hPa], H = 35.20[%rH]
03.May.2020 14:55:19 [location_13321] t = 25.30[°C], p = 995.30[hPa], H = 33.00[%rH]
03.May.2020 14:55:20 [kitchen] t = 32.21[°C], p = 1001.37[hPa], H = 35.20[%rH]
03.May.2020 14:55:20 [location_13321] t = 25.30[°C], p = 995.30[hPa], H = 33.00[%rH]
03.May.2020 14:55:21 [kitchen] t = 32.22[°C], p = 1001.35[hPa], H = 35.20[%rH]
03.May.2020 14:55:21 [location_13321] t = 25.30[°C], p = 995.30[hPa], H = 33.00[%rH]
03.May.2020 14:55:22 [kitchen] t = 32.21[°C], p = 1001.36[hPa], H = 35.20[%rH]
03.May.2020 14:55:22 [location_13321] t = 25.30[°C], p = 995.30[hPa], H = 33.00[%rH]
WARNING: The source is provided as it is without any warranty. Use it on your own risk! The author does not take any responsibility for the damage caused while using this software.
DISCLAIMER: The code is a result of a hobby work and the author is not affiliated with any of the hardware/components/boards manufacturers mentioned in the code, documentation or the description of this project. All trademarks are the property of the respective owners.