Cities are filling up with sensors and smart devices, from air quality stations to street lights and heating in public buildings. Each of them comes with its own protocols and data formats, which makes it hard to build useful applications on top of them. The research in this area covers platforms that connect such devices, process their data in real time and let non-experts decide how the devices should work together.
Much of it has been validated in practice on the Smart City platform of our partner Logimic. Student theses there dealt with storing data from LoRa sensor networks, managing devices and service requests from a phone and visualizing their data on the web 1234567, as well as with controlling street lights, regulating heating, smart home dashboards and planning the routes of service workers 891011. Others made the platform easier to configure, extend and test 1213141516. On the backend, the choice of database turned out to matter: a time series database made the platform faster and cheaper to run 17. Parallel computing then shortened the bulk evaluation of KPIs 1819, and we tested how well machine learning forecasts of energy consumption carry over from one smart home to another 20. A contract project for Logimic in 2023 brought sensors to retention basins to keep track of their water.
Experience from these projects led to two platforms of our own. RIoT processes data from smart devices in real time and monitors their KPIs 21, and it has since been used to monitor road traffic and open data in real time as well 2223. Pocketix lets anyone automate their devices with visual programming, right from a phone 24. Its visual programming language was tested in ACADA, the platform of Logimic, as the video below shows. Theses have since added user-defined functions, a mobile editor, backend processing of the programs and a simulator for evaluating them 25262728, together with applications for managing devices and responsive dashboards for their data 293031.
Data from smart devices matters in many other fields too. With industrial partners, theses processed quality control data of industrial devices, diagnostic data of a bionic arm prosthesis and data of an industrial IoT system 323334; others dealt with beehives, cars and race specials 353637, with organizing the work of gas boiler technicians 38 or with heating holiday homes automatically, a thesis awarded at Excel@FIT and recognized by the rector 39.
Lately the focus has shifted closer to the devices themselves, with a small gateway built on the ESP32 40 and automated testing of edge devices 41 in a self-hosted CI/CD pipeline running on low-power hardware 42, described in more detail on the Dexter project page. Another new topic is the regulation of connected devices: laws such as the EU Cyber Resilience Act, the Radio Equipment Directive or the UK PSTI rules are being turned into machine-readable OSCAL catalogs, so that the compliance of an IoT device can be checked automatically.
