The BMaC project started when sensors were to be added around an office building in order to measure temperature and other parameters, such as relative humidity. After deciding to use ESP8266 MCUs along with DHT22 temperature and humidity sensors, a simple prototype was put together, using a basic firmware written using the Sming framework.
It was found that DHT22 sensors were generally rather bulky and not very precise. The breakout boards used also had an improper resistor mounted on them, leading to the wrong temperature being reported. This sensor type also had the disadvantage of using its own one-wire protocol, instead of a standard interface method.
The DHT22 sensors got swapped out with BME280 MEMS sensors, which measure temperature, humidity, and also air pressure. A CO2 sensor was added as well, in the form of the MH-Z19. This required the firmware to support these additional sensors too. The sensor readings would be sent as MQTT messages, with a backend service subscribing to these topics, and writing them to a time series database (InfluxDB), for viewing and analysis.
Decisions had to be made when the possibility of reading out the counters for products from the fully automatic Jura coffee machines was considered, and with it whether separate firmware would have to be developed.
Instead of separate firmware, the decision was made to use the same firmware for all ESP8266 nodes. This meant that they needed to have the functionality to somehow enable individual features and to support specific sensors and other features. This led to the development of new firmware, which allowed remote commands, sent over MQTT, to toggle feature modules on or off, along with other management features.
Along with the new firmware, a command and control (C&C) server was added, used by the individual nodes to retrieve their configuration, along with an administration application to be used to add new nodes and add or edit the node configuration.
With this framework in place, it became possible to add new features quickly. These included the addition of motion sensors, for detecting the presence of people in a room, to ultimately the controlling of air-conditioning units, as the existing centralized control in the office building was found to be inadequate.
The system as a whole can be visualized like this:
In the upcoming sections, we will be taking a detailed look at each of these aspects.