[go: up one dir, main page]

Skip to content

Indoor environment monitoring on a Raspberry Pi with HomeKit integration. I²C, SPI, UART, GPIO, MQTT. Node.js, Python (and MicroPython!), Arduino, C, C++.

Notifications You must be signed in to change notification settings

aaronhktan/vesta

Repository files navigation

Vesta - Homebridge Plugin Collection

Vesta is the same as the Earth; under both of them is a perpetual fire; the earth and the hearth are symbols of the home. —Ovid, Fasti VI

This repository brings together all the plugins I've written for my indoor environment smart home automation project.

Components

  • homebridge-bme280: Temperature, humidity, barometric air pressure.
  • homebridge-bmp280: Temperature and barometric air pressure.
  • homebridge-dht22: Temperature and humidity.
  • homebridge-micropython-bme280: Temperature, humidity, barometric air pressure from WeMos D1 Mini boards scattered around the house.
  • homebridge-pms7003: Particulate matter, dust, PM2.5/PM10.
  • homebridge-sgp30: Equivalent CO2, volatile organic compounds.
  • homebride-soil-sensor: Analog capacitive soil moisture sensor readings.
  • homebridge-veml6030: White and ambient light brightness.
  • homebridge-display: Displays data on 16x2 character display.

Technical details

Hardware

  • Raspberry Pi! Most of these plugins should work on both the Raspberry Pi 3 and Raspberry Pi Zero. Have not tested on others.
  • The MicroPython port of the BME280 driver runs on a WeMos D1 Mini (ESP8266), and outputs to an SSD1306 I2C display.
  • The soil sensor code runs on either (1) Onion Omega + Arduino Dock, or (2) an ESP8266 running MicroPython.
  • I've added a Fritzing diagram under docs/ to show how everything is wired up.

Protocols

  • I2C, SMBus: SGP30, VEML6030, BME280. These both rely on i2c-dev and ioctl from Linux. The BME280 using MicroPython uses the machine module I2C implementation.
  • SPI: BMP280. This relies on spidev and ioctl from Linux.
  • UART: PMS7003. This relies on termios from Linux. Also uses epoll for asynchronous I/O!
  • GPIO toggling: DHT22. This uses the BCM2835 library.
  • MQTT: Most of the plugins use the MQTT.js library, with the exception of the soil sensor, which relies on paho-mqtt and umqtt for the Onion Omega and MicroPython parts, respectively.

Data flow overview

  • Homebridge runs on the Raspberry Pi, and has access to all the GPIO pins (provided that the homebridge user has been added to the appropriate groups).
  • At a set interval (default 60 seconds, but configurable by plugin), Homebridge polls for data from all the devices connected via hardware.
    • The Node.js plugins invoke the C/Linux API(s) to read data from the hardware modules from JavaScript. This is done via node-bindings (to allow Node.js to talk to the native C++ binding), and node-addon-api (the C++ API that implements the bridge between JavaScript and C APIs).
    • Data that is read is then (1) propagated over MQTT, (2) saved to a local store of for Eve's HomeKit app to consume.
  • Independently, the soil moisture sensors continuously monitor the soil moisture of the pot they're sitting in.
    • In the Omega + Arduino setup, the Arduino dock reads the voltage (built-in 10-bit ADC), and "prints" these over UART for the Omega to consume. The Omega uses a blocking read on the tty port where it is connected; once it receives data, it transmits the reading over MQTT to the Raspberry Pi.
    • On the ESP8266, the 10-bit ADC is limited to a max voltage of 1V, but the NodeMCU board I have has a built-in voltage divider to safely connect the soil moisture sensor. Once it reads a voltage, it transmits the reading over MQTT to the Raspberry Pi.
    • The plugin running on the Raspberry Pi monitors for messages over MQTT. When it receives a reading, averages are calculated and data is stored.
  • Also independently, the BME280 sensors connected to the WeMos D1 Minis are woken up from sleep and polled every 30 seconds by the D1 Mini over I2C, and the read data is sent over MQTT to the Raspberry Pi.
  • Outside of Homebridge, when the Raspberry Pi boots, systemd launches the homebridge-display script. This script subscribes to (some) of the MQTT topics, and displays the data sent on those topics.

Picture

About

Indoor environment monitoring on a Raspberry Pi with HomeKit integration. I²C, SPI, UART, GPIO, MQTT. Node.js, Python (and MicroPython!), Arduino, C, C++.

Topics

Resources

Stars

Watchers

Forks

Languages