The problem of estimating the pollutants in urban areas is one of the most active research in recent years due to the increasing concerns about their influence on human health. Solid state sensors, increasingly small and inexpensive, are being used to build compact multisensor devices. Suffering from sensors instabilities and cross-sensitivities, they need ad-hoc calibration procedures in order to reach satisfying performance levels. In this paper we propose a novel approach based on a Semi Supervised Learning (SSL) system using a Nonlinear AutoRegressive eXogenous model (NARX) to estimate pollutants in urban area and detecting alerts with respect to law limits. We compared our proposal with two other techniques, based on a simple Feed Forward Neural Network and a Semi Supervised Learning FFNN based approach, respectively. Numerical simulations have been carried out to validate the proposed approach on a real dataset.

A new NARX based Semi Supervised Learning algorithm for pollutant estimation

De Vito, S.
2014

Abstract

The problem of estimating the pollutants in urban areas is one of the most active research in recent years due to the increasing concerns about their influence on human health. Solid state sensors, increasingly small and inexpensive, are being used to build compact multisensor devices. Suffering from sensors instabilities and cross-sensitivities, they need ad-hoc calibration procedures in order to reach satisfying performance levels. In this paper we propose a novel approach based on a Semi Supervised Learning (SSL) system using a Nonlinear AutoRegressive eXogenous model (NARX) to estimate pollutants in urban area and detecting alerts with respect to law limits. We compared our proposal with two other techniques, based on a simple Feed Forward Neural Network and a Semi Supervised Learning FFNN based approach, respectively. Numerical simulations have been carried out to validate the proposed approach on a real dataset.
9781479949892
File in questo prodotto:
Non ci sono file associati a questo prodotto.

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

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12079/3545
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
social impact