MP44 - The Dimensionality of MOX Sensors in Air Quality assessment
- Event
- EUROSENSORS XXXVII
2025-09-07 - 2025-09-10
Wroclaw - Band
- Poster I
- Chapter
- Gas Sensors
- Author(s)
- L. Miranda, B. Dorizzi, M. Hariz, D. Petrovska, J. Boudy - Télécom SudParis, Palaiseau (France), N. Redon, C. Duc, M. Verriele - IMT Nord Europe, Douai (France), J. Montalvão - Universidade Federal de Sergipe, São Cristóvão (Brazil)
- Pages
- 291 - 292
- DOI
- 10.5162/EUROSENSORS2025/MP44
- ISBN
- 978-3-910600-07-2
- Price
- free
Abstract
This paper investigates why air quality applications using MOX gas sensors typically rely on sensorarrays with limited number of sensors. Despite their high sensitivity, MOX sensors suffer from low selectivity, andpractical constraints like power and processing capability limit the array size. By analyzing two datasets with differentcomplexities, in terms of monitored pollutants, using three intrinsic dimensionality (ID) estimators, the study showsthat the effective dimensionality of sensor data is often much lower than the total number of sensors. This suggestsredundancy in sensor responses and supports the common use of 2 to 4 MOX sensors. Results also highlight thatfeatures that include some dynamic information, like relative response, yield higher information.