Journal cover Journal topic
Atmospheric Measurement Techniques An interactive open-access journal of the European Geosciences Union
Journal topic

Journal metrics

Journal metrics

  • IF value: 3.248 IF 3.248
  • IF 5-year value: 3.650 IF 5-year
    3.650
  • CiteScore value: 3.37 CiteScore
    3.37
  • SNIP value: 1.253 SNIP 1.253
  • SJR value: 1.869 SJR 1.869
  • IPP value: 3.29 IPP 3.29
  • h5-index value: 47 h5-index 47
  • Scimago H <br class='hide-on-tablet hide-on-mobile'>index value: 60 Scimago H
    index 60
Volume 9, issue 2 | Copyright
Atmos. Meas. Tech., 9, 347-357, 2016
https://doi.org/10.5194/amt-9-347-2016
© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

Research article 04 Feb 2016

Research article | 04 Feb 2016

Mobile sensor network noise reduction and recalibration using a Bayesian network

Y. Xiang, Y. Tang, and W. Zhu Y. Xiang et al.
  • College of Information Engineering, Zhejiang University of Technology, Hangzhou, China

Abstract. People are becoming increasingly interested in mobile air quality sensor network applications. By eliminating the inaccuracies caused by spatial and temporal heterogeneity of pollutant distributions, this method shows great potential for atmospheric research. However, systems based on low-cost air quality sensors often suffer from sensor noise and drift. For the sensing systems to operate stably and reliably in real-world applications, those problems must be addressed. In this work, we exploit the correlation of different types of sensors caused by cross sensitivity to help identify and correct the outlier readings. By employing a Bayesian network based system, we are able to recover the erroneous readings and recalibrate the drifted sensors simultaneously. Our method improves upon the state-of-art Bayesian belief network techniques by incorporating the virtual evidence and adjusting the sensor calibration functions recursively.
Specifically, we have (1) designed a system based on the Bayesian belief network to detect and recover the abnormal readings, (2) developed methods to update the sensor calibration functions infield without requirement of ground truth, and (3) extended the Bayesian network with virtual evidence for infield sensor recalibration. To validate our technique, we have tested our technique with metal oxide sensors measuring NO2, CO, and O3 in a real-world deployment. Compared with the existing Bayesian belief network techniques, results based on our experiment setup demonstrate that our system can reduce error by 34.1% and recover 4 times more data on average.

Publications Copernicus
Download
Short summary
Motivated by unreliable sensor readings and the difficulties in calibrating sensors, we developed a Bayesian-network-based method to remove the abnormal readings and re-calibrate the sensors.
Motivated by unreliable sensor readings and the difficulties in calibrating sensors, we...
Citation
Share