An improved low-power measurement of ambient NO<sub>2</sub> and O<sub>3</sub> combining electrochemical sensor clusters and machine learning
<p>Low-cost sensors (LCSs) are an appealing solution to the problem of spatial resolution in air quality measurement, but they currently do not have the same analytical performance as regulatory reference methods. Individual sensors can be susceptible to analytical cross-interferences; have ra...
Main Authors: | , , , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Copernicus Publications
2019-02-01
|
Series: | Atmospheric Measurement Techniques |
Online Access: | https://www.atmos-meas-tech.net/12/1325/2019/amt-12-1325-2019.pdf |
_version_ | 1811240318876516352 |
---|---|
author | K. R. Smith P. M. Edwards P. D. Ivatt J. D. Lee J. D. Lee F. Squires C. Dai R. E. Peltier M. J. Evans M. J. Evans Y. Sun A. C. Lewis A. C. Lewis |
author_facet | K. R. Smith P. M. Edwards P. D. Ivatt J. D. Lee J. D. Lee F. Squires C. Dai R. E. Peltier M. J. Evans M. J. Evans Y. Sun A. C. Lewis A. C. Lewis |
author_sort | K. R. Smith |
collection | DOAJ |
description | <p>Low-cost sensors (LCSs) are an appealing solution to the problem of spatial
resolution in air quality measurement, but they currently do not have the
same analytical performance as regulatory reference methods. Individual
sensors can be susceptible to analytical cross-interferences; have random
signal variability; and experience drift over short, medium and long
timescales. To overcome some of the performance limitations of individual
sensors we use a clustering approach using the instantaneous median signal
from six identical electrochemical sensors to minimize the randomized drifts
and inter-sensor differences. We report here on a low-power analytical device
(<span class="inline-formula"><i><</i> 200</span> W) that is comprised of clusters of sensors for
<span class="inline-formula">NO<sub>2</sub></span>, <span class="inline-formula">O<sub><i>x</i></sub></span>, CO and total volatile organic compounds
(VOCs) and that measures supporting parameters such as water vapour and temperature.
This was tested in the field against reference monitors, collecting ambient
air pollution data in Beijing, China. Comparisons were made of <span class="inline-formula">NO<sub>2</sub></span>
and <span class="inline-formula">O<sub><i>x</i></sub></span> clustered sensor data against reference methods for
calibrations derived from factory settings, in-field simple linear regression
(SLR) and then against three machine learning (ML) algorithms. The parametric
supervised ML algorithms, boosted regression trees (BRTs) and boosted linear
regression (BLR), and the non-parametric technique, Gaussian process (GP),
used all available sensor data to improve the measurement estimate of
<span class="inline-formula">NO<sub>2</sub></span> and <span class="inline-formula">O<sub><i>x</i></sub></span>. In all cases ML produced an
observational value that was closer to reference measurements than SLR alone.
In combination, sensor clustering and ML generated sensor data of a quality
that was close to that of regulatory measurements (using the RMSE metric) yet
retained a very substantial cost and power advantage.</p> |
first_indexed | 2024-04-12T13:17:03Z |
format | Article |
id | doaj.art-1155447e166c4d859e9d4caca0343a3d |
institution | Directory Open Access Journal |
issn | 1867-1381 1867-8548 |
language | English |
last_indexed | 2024-04-12T13:17:03Z |
publishDate | 2019-02-01 |
publisher | Copernicus Publications |
record_format | Article |
series | Atmospheric Measurement Techniques |
spelling | doaj.art-1155447e166c4d859e9d4caca0343a3d2022-12-22T03:31:38ZengCopernicus PublicationsAtmospheric Measurement Techniques1867-13811867-85482019-02-01121325133610.5194/amt-12-1325-2019An improved low-power measurement of ambient NO<sub>2</sub> and O<sub>3</sub> combining electrochemical sensor clusters and machine learningK. R. Smith0P. M. Edwards1P. D. Ivatt2J. D. Lee3J. D. Lee4F. Squires5C. Dai6R. E. Peltier7M. J. Evans8M. J. Evans9Y. Sun10A. C. Lewis11A. C. Lewis12Wolfson Atmospheric Chemistry Laboratories, University of York, York, YO10 5DD, UKWolfson Atmospheric Chemistry Laboratories, University of York, York, YO10 5DD, UKWolfson Atmospheric Chemistry Laboratories, University of York, York, YO10 5DD, UKWolfson Atmospheric Chemistry Laboratories, University of York, York, YO10 5DD, UKNational Centre for Atmospheric Science, University of York, York, YO10 5DD, UKWolfson Atmospheric Chemistry Laboratories, University of York, York, YO10 5DD, UKWolfson Atmospheric Chemistry Laboratories, University of York, York, YO10 5DD, UKEnvironmental Health Science, University of Massachusetts, 686 North Pleasant Street, Amherst, MA 01003, USAWolfson Atmospheric Chemistry Laboratories, University of York, York, YO10 5DD, UKNational Centre for Atmospheric Science, University of York, York, YO10 5DD, UKState Key Laboratory of Atmospheric Boundary Layer Physics and Atmospheric Chemistry, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, ChinaWolfson Atmospheric Chemistry Laboratories, University of York, York, YO10 5DD, UKNational Centre for Atmospheric Science, University of York, York, YO10 5DD, UK<p>Low-cost sensors (LCSs) are an appealing solution to the problem of spatial resolution in air quality measurement, but they currently do not have the same analytical performance as regulatory reference methods. Individual sensors can be susceptible to analytical cross-interferences; have random signal variability; and experience drift over short, medium and long timescales. To overcome some of the performance limitations of individual sensors we use a clustering approach using the instantaneous median signal from six identical electrochemical sensors to minimize the randomized drifts and inter-sensor differences. We report here on a low-power analytical device (<span class="inline-formula"><i><</i> 200</span> W) that is comprised of clusters of sensors for <span class="inline-formula">NO<sub>2</sub></span>, <span class="inline-formula">O<sub><i>x</i></sub></span>, CO and total volatile organic compounds (VOCs) and that measures supporting parameters such as water vapour and temperature. This was tested in the field against reference monitors, collecting ambient air pollution data in Beijing, China. Comparisons were made of <span class="inline-formula">NO<sub>2</sub></span> and <span class="inline-formula">O<sub><i>x</i></sub></span> clustered sensor data against reference methods for calibrations derived from factory settings, in-field simple linear regression (SLR) and then against three machine learning (ML) algorithms. The parametric supervised ML algorithms, boosted regression trees (BRTs) and boosted linear regression (BLR), and the non-parametric technique, Gaussian process (GP), used all available sensor data to improve the measurement estimate of <span class="inline-formula">NO<sub>2</sub></span> and <span class="inline-formula">O<sub><i>x</i></sub></span>. In all cases ML produced an observational value that was closer to reference measurements than SLR alone. In combination, sensor clustering and ML generated sensor data of a quality that was close to that of regulatory measurements (using the RMSE metric) yet retained a very substantial cost and power advantage.</p>https://www.atmos-meas-tech.net/12/1325/2019/amt-12-1325-2019.pdf |
spellingShingle | K. R. Smith P. M. Edwards P. D. Ivatt J. D. Lee J. D. Lee F. Squires C. Dai R. E. Peltier M. J. Evans M. J. Evans Y. Sun A. C. Lewis A. C. Lewis An improved low-power measurement of ambient NO<sub>2</sub> and O<sub>3</sub> combining electrochemical sensor clusters and machine learning Atmospheric Measurement Techniques |
title | An improved low-power measurement of ambient NO<sub>2</sub> and O<sub>3</sub> combining electrochemical sensor clusters and machine learning |
title_full | An improved low-power measurement of ambient NO<sub>2</sub> and O<sub>3</sub> combining electrochemical sensor clusters and machine learning |
title_fullStr | An improved low-power measurement of ambient NO<sub>2</sub> and O<sub>3</sub> combining electrochemical sensor clusters and machine learning |
title_full_unstemmed | An improved low-power measurement of ambient NO<sub>2</sub> and O<sub>3</sub> combining electrochemical sensor clusters and machine learning |
title_short | An improved low-power measurement of ambient NO<sub>2</sub> and O<sub>3</sub> combining electrochemical sensor clusters and machine learning |
title_sort | improved low power measurement of ambient no sub 2 sub and o sub 3 sub combining electrochemical sensor clusters and machine learning |
url | https://www.atmos-meas-tech.net/12/1325/2019/amt-12-1325-2019.pdf |
work_keys_str_mv | AT krsmith animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT pmedwards animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT pdivatt animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT jdlee animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT jdlee animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT fsquires animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT cdai animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT repeltier animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT mjevans animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT mjevans animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT ysun animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT aclewis animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT aclewis animprovedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT krsmith improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT pmedwards improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT pdivatt improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT jdlee improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT jdlee improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT fsquires improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT cdai improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT repeltier improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT mjevans improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT mjevans improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT ysun improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT aclewis improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning AT aclewis improvedlowpowermeasurementofambientnosub2subandosub3subcombiningelectrochemicalsensorclustersandmachinelearning |