HEPTA (Helmholtz European Partnership for Technological Advancement) – Topic area 3: Smart Cities

February 23rd, 2023  |  Published in Research

HEPTA promotes the cooperation between the Aristotle University of Thessaloniki (AUTh) and KIT in the development of sustainable technologies in the areas of “air quality”, “atmospheric physics”, “biomass” and “smart cities”. HEPTA has three goals: First, a formal connection between AUTh and KIT for technology development is to be established with special consideration of the most important research topics in the fields of climate, energy, and environment. Secondly, the common scientific offspring – with a strictly equal proportion of women and men – should be promoted. Thirdly, the project partners want to further expand the cooperation between the two institutions with a long-term perspective in order to transfer research results more quickly into concrete applications together with industry.

Topic area 3: Smart Cities has the following focuses:

  • Surrogate / Inverse modeling of physical phenomena using Machine Learning focusing on aerosol and gas distribution in cities
  • Using Gaussian Processes and equivalent deep and bayesian neural network Models for modelling such spatiotemporal phenomena
  • Fitting (land-use-) regression models driven by stationary and mobile low-cost (uncalibrated, noisy) sensor input and other heterogeneous static and dynamic information sources (like traffic sensors, satellite images, news reports) related to cities and urban mobility.


  • Li, C.; Budde, M.; Tremper, P.; Schäfer, K.; Riesterer, J.; Redelstein, J.; Petersen, E.; Khedr, M.; Liu, X.; Köpke, M.; Hussain, S.; Ernst, F.; Kowalski, M.; Pesch, M.; Werhahn, J.; Hank, M.; Philipp, A.; Cyrys, J.; Schnelle-Kreis, J.; Grimm, H.; Ziegler, V.; Peters, A.; Emeis, S.; Riedel, T.; Beigl, M. (2022) SmartAQnet 2020: A New Open Urban Air Quality Dataset from Heterogeneous PM Sensors. ProScience, 8. doi:10.14644/dust2021.001
  • Li, C., Riedel, T., & Beigl, M. (2022). Neural Kernel Network Deep Kernel Learning for Predicting Particulate Matter from Heterogeneous Sensors with Uncertainty. In International Conference on Information Integration and Web (pp. 252-266). Springer, Cham.

Comments are closed.