%0 Conference Proceedings %T Source term estimation: variational method versus machine learning applied to urban air pollution %+ Département Systèmes (DSYS) %+ Université Grenoble Alpes (UGA) %A Lopez-Ferber, Roman %A Leirens, Sylvain %A Georges, Didier %< avec comité de lecture %B CSC 2022 - IFAC Workshop on Control for Smart Cities %C Sozopol (virtual), Bulgaria %8 2022-06-27 %D 2022 %K Source Detection %K Source Term Estimation %K variational methods %K 3D-Var %K air pollution %K neural network %K advection-diffusion %Z Computer Science [cs]/Signal and Image Processing %Z Statistics [stat]/Machine Learning [stat.ML] %Z Environmental SciencesConference papers %X Source detection is a field of study gaining interest due to environmental concerns about air quality in populated areas. We developed a machine learning framework inspired by previous works on road traffic estimation, and compared it to a classical variational method under a unidimensional and stationary problem. We tested source reconstruction with datasets coming from 12 and 50 sensors with and without noise. Noise was set to follow a gaussian law with a dependent variance from the maximum measured value of a concentration profile. Both methods are reasonably robust to noise. The results reveal that the Neural Network used here, a multilayer perceptron, performs very well compared to the classical 3D-Var method. %G English %2 https://cea.hal.science/cea-03716399/document %2 https://cea.hal.science/cea-03716399/file/IFAC_8_Nov_2021.pdf %L cea-03716399 %U https://cea.hal.science/cea-03716399 %~ SDE %~ CEA %~ UGA %~ GIP-BE %~ DRT %~ LETI %~ CEA-GRE %~ UGA-EPE %~ TEST3-HALCNRS %~ MAP-CEA