Forecasting future values of air quality related metrics and specific pollutant concentration could be of pivotal importance in recent Smart City perspectives. A number of pollutants are dangerous for people’s health and impact on environment and climate. In order to control and reduce the emissions, national and international organizations have defined guidelines and targeted limits to be respected currently, and to be progressively reduced along the year/months. On this regard, the European Union has set limits for the concentration of the yearly mean value of NO2 which must not exceed 40 µg/m3. To this end, in this paper, we propose a model and tool to compute long terms predictions, up to 180 days in advance, of the progressive mean value of NO2 with a precision needed to enable decision makers to perform corrections. The solution proposed is based on machine learning approach taking into account measures of pollutant, traffic flow, weather and environmental variables coming from sensors on the field. A comparison of different techniques has been provided. The research activity has been developed in the context to TRAFAIR CEF project of EC which aimed to study the effect of traffic and of other human activities on NO and NO2. The data and the solution have been developed by exploiting the Snap4City platform; the validation of the solution has been performed by using actual measured data from years 2014 to 2020 in the area of Florence, Italy. The results are accessible via a monitoring dashboard on Snap4City which reports real time values and predictions in real time.

Long Term Predictions of NO2 Average Values via Deep Learning / Bellini P.; Bilotta S.; Cenni D.; Collini E.; Nesi P.; Pantaleo G.; Paolucci M.. - STAMPA. - 12956 LNCS:(2021), pp. 595-610. (Intervento presentato al convegno International Conference on Computational Science and Its Applications - ICCSA 2021 tenutosi a Cagliari, Italy nel September 13–16, 2021) [10.1007/978-3-030-87010-2_44].

Long Term Predictions of NO2 Average Values via Deep Learning

Bellini P.;Bilotta S.;Cenni D.;Collini E.;Nesi P.
;
Pantaleo G.;Paolucci M.
2021

Abstract

Forecasting future values of air quality related metrics and specific pollutant concentration could be of pivotal importance in recent Smart City perspectives. A number of pollutants are dangerous for people’s health and impact on environment and climate. In order to control and reduce the emissions, national and international organizations have defined guidelines and targeted limits to be respected currently, and to be progressively reduced along the year/months. On this regard, the European Union has set limits for the concentration of the yearly mean value of NO2 which must not exceed 40 µg/m3. To this end, in this paper, we propose a model and tool to compute long terms predictions, up to 180 days in advance, of the progressive mean value of NO2 with a precision needed to enable decision makers to perform corrections. The solution proposed is based on machine learning approach taking into account measures of pollutant, traffic flow, weather and environmental variables coming from sensors on the field. A comparison of different techniques has been provided. The research activity has been developed in the context to TRAFAIR CEF project of EC which aimed to study the effect of traffic and of other human activities on NO and NO2. The data and the solution have been developed by exploiting the Snap4City platform; the validation of the solution has been performed by using actual measured data from years 2014 to 2020 in the area of Florence, Italy. The results are accessible via a monitoring dashboard on Snap4City which reports real time values and predictions in real time.
2021
Computational Science and Its Applications - ICCSA 2021
International Conference on Computational Science and Its Applications - ICCSA 2021
Cagliari, Italy
September 13–16, 2021
Bellini P.; Bilotta S.; Cenni D.; Collini E.; Nesi P.; Pantaleo G.; Paolucci M.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificatore per citare o creare un link a questa risorsa: https://hdl.handle.net/2158/1289946
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 2
  • ???jsp.display-item.citation.isi??? 1
social impact