Weather stations


Julian, Hannes, Jochen

LUBW weather station at FZI

Labels: Stream Processing, Batch Processing

  • the device was mounted at FZI in January 2017 for measuring air pressure, temperature, wind speed, precipitation (was bedeuten die Parameter "strg_1" und "rlf_1").
  • measurement results (*.mw1.xml) are pushed via xml-files on BigGIS SFTP Server at the University Konstanz -measurement devices will be dismounted at April 2018

SenseBox-based weather stations

  • support two different modes of data transmission: WLAN and LoRa

    • LoRa is preferred for longer range and easier deployment (no site-specific configuration)
  • LoRa gateways

    • Two LoRa gateways should be deployed, one at FZI, the other at LUBW-KA (TODO:clarify)
    • Gateways are a Kickstarter project, and have not yet been delivered as of 2017-12-15. Last status on Kickstarter is that the gateways are ready to ship (2017-11-06)
    • Julian handles Kickstarter and deployment
  • Data handling

    • a common Kafka queue for events from both transports
    • one Kafka message per transmitted event, containing all measurements (temperature, humidity, air pressure, internal temperature, light, UV radiation)
    • transport specific adapters for WLAN and LoRa
    • should be handled in a stream-processing way (pipeline modeled using StreamPipes)
    • should be stored within BigGIS database i.e. Accumulo/Exasolution (sensor,lat,lon,ts)
      • a Flink job that persists the events from the Kafka queue into MySQL or PostgreSQL exists, could be extended to support other JDBC databases
    • should be sent to (TODO:Jochen)
    • Outlier filtering node (kafka->flink->kafka)
  • Deployment

    • 34 LoRa sensor units should be deployed, Julian handles locations, external organizations etc.

Web-based mobile-friendly app

  • QR code contains stations id and URL that leads to public web
    • the web page contains info about the station and the project
    • admin can click and change station information (or register a new station)
    • lat/lon is taken from the phone (HTML5 geolocation api)
    • admin can add additional parameters (placement details)