1ec25e9f-ab2f-4535-8474-63d3a6a7d52e Andrea Vianello Eurac Research - Center for Sensing Solutions Researcher Viale Druso, 1 / Drususallee 1, eurac research Bolzano Autonomous Province of Bolzano 39100 Italy supportcss@eurac.edu http://www.eurac.edu/en/research/mountains/remsen/Pages/default.aspx WWW:LINK-1.0-http--link remote sensing eurac research 2026-05-07T08:33:12.053Z ISO 19139 1.0 Air Quality Utilizing Indoor AI Sensors (AQUINAS) - Input data AQUINAS 2026-03-27T00:00:00+01:00 2026-03-27T00:00:00+01:00 2025 publisher https://www.eurac.edu/en 10.48784/1ec25e9f-ab2f-4535-8474-63d3a6a7d52e Mendicino, R., & Hussain, S. F. (2026). Air Quality Utilizing Indoor AI Sensors (AQUINAS) - Input data (Version 2025) [Data set]. Center for Sensing Solutions. https://doi.org/10.48784/1EC25E9F-AB2F-4535-8474-63D3A6A7D52E This research provides knowledge on the performance and limitations of low-cost environmental sensors in real usage conditions. The EQ-OX platform will be used and the collected data will be compared to those of the reference station. The main purpose of this research is to compare these data and try to correct them through the use of machine learning algorithms. Project focused on the evaluation and correction of low-cost indoor and outdoor air quality sensors using machine learning, with comparison against reference stations. Roberto Mendicino Eurac Research - Center for Sensing Solutions roberto.mendicino@eurac.edu Simone Tritini Eurac Research - Center for Sensing Solutions simone.tritini@eurac.edu Roberto Mendicino Eurac Research - Center for Sensing Solutions roberto.mendicino@eurac.edu https://edp-portal.eurac.edu/geonetwork/srv/api/records/1ec25e9f-ab2f-4535-8474-63d3a6a7d52e/attachments/aquinas_input_data.png InfluxDB 1.8 GEMET - INSPIRE themes, version 1.0 2018-05-23T10:25:56 geonetwork.thesaurus.external.theme.httpinspireeceuropaeutheme-theme air quality low-cost sensors temperature humidity database Attribution 4.0 International (CC BY 4.0): You are free to share and adapt under the following conditions: attribution and no additional restriction. (https://creativecommons.org/licenses/by/4.0/deed.en) 2024-02-13 2025-07-11 23.5935 23.8622 37.878 38.0895 Monitoring indoor and outdoor air quality (ΑQ), including pollutants and thermo-hygrometric parameters, is a key component of any strategy to improve population health and wellbeing. The WHO estimated 7.2 million deaths globally in 2019 due to air pollution, with 46% of this mortality being due to indoor exposure. Another fact that goes under the radar, and denotes the importance of household air pollution, is that the WHO AQ guidelines are equally applicable for ambient and indoor exposure. However, the attention diverted to indoor AQ characterization is limited, especially when considering the length of indoor exposure but also how difficult it would be to regulate indoor emissions. To improve the reliability and facilitate the spread of EQ-OX as a platform that provides long-term data at low cost, a calibration algorithm based on artificial intelligence (AI) for the adjustment of sensor outputs towards reference-grade measurements is being developed, with encouraging results. A salient aspect of the calibration procedure of low-cost sensors involves their comparison with reference instruments or pre-calibrated sensors, in controlled experiments. However, in actual atmospheric conditions, low-cost sensors are not very selective and cross-links between atmospheric properties can be observed, leading to sensors underperforming. In this exploratory study, the EQ-OX sensors will be calibrated in ambient conditions and in parallel, they will be applied to obtain indicative indoor measurements. The main objectives of this research are: • Compare the low-cost sensors integrated in EQ-OX to calibrated reference instruments across different ambient conditions at ATMOS-NOA • Calibrate the EQ-OX sensors using classic statistical methods and showcase their limitations and implementing the “long short-term memory” algorithm trying to increase the sensitivity and precision combining the outputs of multiple sensors. • Evaluate the performance of EQ-OX through comparison with reference instruments in different pollution and meteorological conditions and test their operation in actual indoor spaces ATMOS_NOA is selected as the host facility as it has all the necessary instrumentation for the calibration experiments. It is in an area with diverse anthropogenic and natural sources, while the meteorological conditions with strong mesoscale dynamics and an oxidizing atmosphere promote secondary pollutant formation and transport. Moreover, the host’s personnel are experienced in the calibration of low-cost sensors and the use of portable instruments for personal exposure assessment. This research was conducted because it is of strategic interest for the Center for Sensing Solution. In particular, the use of artificial intelligence algorithms to improve measurements based on low-cost sensors is a fundamental element in the research activity of the center. Furthermore, thanks to the collaboration with the Renewable Energy Group, it was possible to utilize and try to optimize the sensor node developed for indoor air quality. We hope for future collaboration with the NOA group from Athens to continue research on low-cost sensors for air quality monitoring. Text, table Center for Sensing Solutions Viale Druso 1 Bolzano South-Tyrol 39100 Italy supportcss@eurac.edu http://saocompute.eurac.edu/sensordb WWW:LINK-1.0-http--link ATMO - API InfluxDB Access restricted. To access the data, please contact the persons listed in "Point of contact" section. https://doi.org/10.48784/1ec25e9f-ab2f-4535-8474-63d3a6a7d52e DOI Digital Object Identifier (DOI) The characterization of indoor air is performed mostly at citizen science level and it is usually limited to few parameters. Eurac Research has recently developed the Environmental Quality bOX (EQ-OX) multi-sensor device. It is a low-cost system with different sensors for a wide array of parameters relevant for the indoor environment, including CO, CO2, NO2, O3, PMs and TVOC.