Multi-unit calibration rejects inherent device variability of chemical sensor arrays

• Published in: Sensors and actuators. B, Chemical Vol. 265; pp. 142 - 154
• Main Authors: Solórzano, Ana, Rodríguez-Pérez, Raquel, Padilla, Marta, Graunke, Thorsten, Fernandez, Luis, Marco, Santiago, Fonollosa, Jordi
• Format: Journal Article Publication
• Language: English
• Published: Lausanne Elsevier B.V 15.07.2018; Elsevier Science Ltd
• Subjects: Calibration; Calibration transfer; Centrifuges; Chemical sensors; Classification; Detectors de gasos; Enginyeria electrònica; Gas sensor array; Gas sensors; Instrumentació i mesura; Machine olfaction; Mathematical models; Model accuracy; Model testing; MOX; MOX sensor; Nanoscience; Nanostructured materials; Organic chemistry; Particle accelerators; Performance prediction; Robust calibration; Sensor arrays; Sensors; Sensors i actuadors; Sensors químics; Superconductivity; Àrees temàtiques de la UPC; MOX sensor; Calibration transfer; Gas sensor array; Machine olfaction; Robust calibration
• ISSN: 0925-4005; 1873-3077
• DOI: 10.1016/j.snb.2018.02.188

•We present a calibration strategy that uses multiple copies of a gas sensor array.•Obtained calibration models can be extended to new copies of the system without significant drop in the prediction accuracy.•Our strategy does not require the acquisition of calibration transfer samples.•Our methodology results in a cost-efficient calibration strategy, favoring mass-production applications for gas sensor arrays. Inherent sensor variability limits mass-production applications for metal oxide (MOX) gas sensor arrays because calibration for replicas of a sensor array needs to be performed individually. Recently, calibration transfer strategies have been proposed to alleviate calibration costs of new replicas, but they still require the acquisition of transfer samples. In this work, we present calibration models that can be extended to uncalibrated replicas of sensor arrays without acquiring new samples, i.e., general or global calibration models. The developed methodology consists in including multiple replicas of a sensor array in the calibration process such that sensor variability is rejected by the general model. Our approach was tested using replicas of a MOX sensor array in the classification task of six gases and synthetic air, presented at different background humidity and concentration levels. Results showed that direct transfer of individual calibration models provides poor classification accuracy. However, we also found that general calibration models kept predictive performance when were applied directly to new copies of the sensor array. Moreover, we explored, through feature selection, whether particular combinations of sensors and operating temperatures can provide predictive performances equivalent to the calibration model with the complete array, favoring thereby the existence of more robust calibration models.