Javier González Monroy, Achim J. Lilienthal, Jose Luis Blanco, Javier González Jimenez and Marco Trincavelli
Probabilistic Gas Quantification with MOX Sensors in Open Sampling Systems - A Gaussian Process Approach
Sensors and Actuators B: Chemical, Volume 188, Issue 0, November 2013, pp. 298 - 312.
Abstract: Gas quantification based on the response of an array of metal oxide (MOX) gas sensors in an open sampling system is a complex problem due to the highly dynamic characteristic of turbulent airflow and the slow dynamics of the MOX sensors. However, many gas related applications require to deter- mine the gas concentration the sensors are being exposed to. Due to the chaotic nature that dominates gas dispersal, in most cases it is desirable to provide, together with an estimate of the mean concentration, an estimate of the uncertainty of the prediction. This work presents a probabilistic ap- proach for gas quantification with an array of MOX gas sensors based on Gaussian Processes, estimating for every measurement of the sensors a pos- terior distribution of the concentration, from which confidence intervals can be obtained. The proposed approach has been tested with an experimental setup where an array of MOX sensors and a Photo Ionization Detector (PID), used to obtain ground truth concentration, are placed downwind with respect to the gas source. Our approach has been implemented and compared with standard gas quantification methods, demonstrating the advantages when estimating gas concentrations.
Keywords: Gas Quantification, Open Sampling System, MOX sensors, Gaussian Processes
Download:
Paper: [PDF (14.3MB)]
@ARTICLE{Vergara_etal:SNB:2013,
  AUTHOR = {González Monroy, Javier and Lilienthal, Achim and Blanco, Jose Luis and González Jimenez, Javier and Trincavelli, Marco},
  TITLE = {Probabilistic gas quantification with \{MOX\} sensors in Open Sampling Systems - A Gaussian Process approach},
  JOURNAL = {Sensors and Actuators B: Chemical},
  YEAR = {2013},
  VOLUME = {188},
  ISSUE = {0},
  PAGES = {298 - 312}
  DOI = {http://dx.doi.org/10.1016/j.snb.2013.06.053}
}