Sensors & Air Quality

national
bioengineering
air-quality
Electronic noses, CO₂ monitoring, wearable personal exposure monitors, and indoor air quality interventions — B2SLab’s sensor programme spanning food quality, school environments, asthma management, and global health.
Keywords

electronic nose, air quality, CO2, sensors, asthma, wearable, indoor air quality, Nepal, schools, Jordi Fonollosa, UPC, bioengineering

Funding Ministerio de Ciencia e Innovación (AEI) · UPC Social Commitment
Status Active (multiple ongoing lines)
B2SLab PI Jordi Fonollosa

Overview

Air quality shapes health in ways that are often invisible — both literally, because most pollutants cannot be seen, and analytically, because the tools required to measure personal exposure have historically been expensive, bulky, and impractical for real-world deployment.

B2SLab’s sensor programme addresses this gap through a decade-long research effort covering four distinct but connected areas: electronic noses for chemical sensing and food quality; CO₂ monitoring in occupied indoor spaces; wearable personal exposure monitors for vulnerable patient groups; and biomass combustion and indoor air quality in low-resource settings. The programme develops sensor hardware, signal processing algorithms, machine learning models, and the deployment strategies needed to make these technologies work in demanding real environments.

This sensor expertise also directly underpins B2SLab’s contribution to the SERVICO2 project, where the group is responsible for the multi-sensor nodes deployed at headwater catchment sites across Europe to monitor greenhouse gas fluxes.


CO₂ monitoring in school environments

The placement problem

CO₂ concentration is the most widely used proxy for indoor air quality and ventilation adequacy in occupied spaces. During and after the COVID-19 pandemic, thousands of schools in Catalonia and Spain installed CO₂ monitors to guide ventilation decisions. The implicit assumption was that a single sensor per classroom could reliably characterise the air quality of the whole room.

It cannot.

Note

Key finding — Indoor Air, 2025

B2SLab’s study of CO₂ spatial distribution within classrooms found that sensor placement alone can produce discrepancies exceeding 100 ppm between readings taken simultaneously in the same room. These differences are large enough to trigger or suppress ventilation alerts incorrectly — meaning that the sensor may be indicating safe air quality while part of the occupied space is significantly under-ventilated, or vice versa.

The study characterised how room geometry, occupant distribution, heating/cooling airflow, and sensor height interact to produce these spatial gradients, and derived evidence-based placement guidelines for single-sensor deployments.

These results have direct practical implications for school ventilation policies across Catalonia, Spain, and Europe — most of which currently offer no guidance on sensor placement.


Wearable air quality monitors for asthma patients

Personal exposure, not building-level averages

Population-level air quality monitoring — regulatory stations, school CO₂ sensors, building ventilation alarms — characterises the environment but not the individual. A child with asthma does not experience the average air quality of their municipality; they experience the specific pollutant concentrations in the spaces where they spend their time, in the specific temporal sequence in which they encounter them.

Wearable personal exposure monitors address this by continuously measuring what the individual actually breathes — enabling, for the first time, direct linkage between a patient’s pollutant exposure trajectory and their respiratory symptom diary.

Tip

Key finding — Environment International, 2026

A 6-month personal exposure monitoring study in 13 asthma patients used miniaturised wearable sensors to track NO₂, PM₂.₅, CO₂, temperature, and humidity continuously throughout daily life.

Patients’ personal exposures showed substantial variation between individuals and over time, with exposure peaks — in traffic, cooking, and cleaning contexts — that were systematically missed by the nearest regulatory monitoring station. Personal exposure profiles were associated with self-reported symptom severity, providing direct evidence for the link between pollutant micro-exposure and asthma control that regulatory monitoring cannot establish.

This work supports the development of closed-loop sensor tools that can identify a patient’s high-exposure contexts and help clinicians tailor advice — not based on where the patient lives, but on what they actually breathe.


Electronic noses: chemical sensing and signal processing

Gas sensor arrays for complex mixtures

An electronic nose (e-nose) is an array of non-specific gas sensors whose combined response pattern encodes information about the chemical composition of a gas sample. Unlike a mass spectrometer, an e-nose cannot directly identify individual compounds; its power lies in pattern recognition — distinguishing complex gas mixtures, tracking changes over time, and classifying samples without requiring identification of every constituent.

B2SLab has developed signal processing and machine learning methods for e-nose systems across a range of application domains:

Note

E-nose applications at B2SLab

  • Food quality control — classification of food freshness, origin, and adulteration using sensor array responses to volatile organic compound mixtures
  • Home monitoring for elderly — gas-sensor-based detection of behavioural anomalies (missed meals, incomplete cooking, unusual patterns) as a non-invasive assisted-living tool
  • Environmental monitoring — gas sensor arrays for outdoor deployment in challenging conditions, including the multi-sensor nodes used in SERVICO2

A persistent challenge in e-nose systems is instrument variability: sensors drift over time, and calibration models trained on one instrument degrade when applied to another. B2SLab has developed multivariate preprocessing and calibration transfer methods for MCC-IMS (Multicapillary Column Ion Mobility Spectrometry) based e-nose systems that substantially reduce this degradation.

Published: Sensors, 2025.


Nepal indoor air quality: biomass combustion and global health

Indoor air pollution as a neglected health burden

In high-income countries, indoor air quality concerns centre on NO₂, PM₂.₅, and VOC from traffic, cooking, and building materials. In much of the world, the dominant indoor air quality hazard is biomass combustion — the use of wood, dung, or agricultural residues for cooking and heating in poorly ventilated homes. The resulting indoor PM concentrations can be orders of magnitude above WHO guidelines, and the health consequences — respiratory disease, cardiovascular disease, adverse birth outcomes — disproportionately affect women and young children who spend the most time near cooking fires.

Tip

Award-winning intervention — UPC Social Commitment Prize 2025

B2SLab participated in a project in Nepal that evaluated the impact of an improved chimney stove intervention on indoor particulate matter concentrations in rural households. Before the intervention, kitchens with open biomass fires had PM concentrations vastly exceeding safe thresholds.

After installation of the improved stoves — which direct combustion gases through a chimney rather than into the living space — indoor PM concentrations were reduced by approximately 90%.

The project received the UPC Social Commitment Prize 2025, recognising it as an outstanding example of university research contributing directly to health equity and sustainable development goals.

This work exemplifies the reach of sensor methodology when it is deployed in service of global health challenges rather than solely in high-resource research settings.


Connection to SERVICO2

The sensor expertise developed across these four research lines now forms the technical backbone of B2SLab’s contribution to SERVICO2 — a European Water4All project investigating greenhouse gas regulation in headwater catchments.

For SERVICO2, the group is designing, building, and deploying multi-sensor monitoring nodes capable of measuring CO₂, CH₄, N₂O, temperature, humidity, and VOCs in outdoor riparian environments — requiring all of the competencies developed in the indoor and wearable sensor lines: gas sensor selection, baseline drift correction, cross-sensitivity compensation, multivariate calibration, and rugged low-power hardware design for remote deployment.