Delta T Sensor Node

This experimental project is about applying clean architecture to an IoT code base. Not because it's easy but because it's hard ;).

The hardware setup is quite simple: An ESP8266, a DHT22 sensor and an Adafruit SSD1306 OLED display.

Every few seconds the firmware on the ESP8266 reads temperature and relative humidity from the attached DHT22 sensor, shows the data on the display and publishes the data via MQTT on a temperature and a relative humidity topic.

Architecture

Code

lib/facades

Convenience code to execute the "Publish Sensor Data" use-case.

lib/drivers

Code to access hardware and network, e.g. DHT22 sensor, Serial line, SSD1306 display, MQTT message queue, and WLAN network interface (blue layer in the clean architecture figure).

lib/interface_adapters

Code for "Publish Sensor Data" presenter, view-model, and controller (green layer in the clean architecture figure) and gateway APIs.

lib/application_business_rules

Code for "Publish Sensor Data" interactor (red layer in the clean architecture figure).

Actually there are no enterprise business rules.

Class Diagram

The class color code corresponds to the colors of the clean architecture layers.

Example Deployment

In this example the Delta T Sensor Node (delta-t-sensor-node) and Dekta T Alarm Node (delta-t-alarm-node) firmwares are deployed on ESP8266 microchips. The Delta T Collector (delta-t-collector) and the Delta T Server (delta-t-server) components are deployed on a Raspberry Pi, where the

• delta-t-sensor-node firmware reads temperature and humidity data from a DHT22 sensor and publishes the data on MQTT message queue topics
• delta-t-collector component subscribes on MQTT message queue topics, collects data and forwards them via REST API
• delta-t-server component gets sensor data via REST API and starts/stops alarm via REST API
• delta-t-alarm-node firmware gets commands to start/stop alarms via REST API

Build

Copy include/config.h.in to include/config.h and edit it to your needs. Then start the build process, e.g. by executing pio run in the project directory.