Air pollution is a major public health challenge, yet high-quality air quality data remain spatially limited due to the high costs of reference-grade monitoring stations. AIR4LIFE develops and validates a low-cost, open-source air quality monitoring framework based on IoT and AI technologies to enable high-resolution urban air quality monitoring. The project integrates affordable IoT sensors with AI-based data quality assurance to deliver reliable, scalable, and transparent environmental health data. Pilot deployments are carried out in Bremen, Bochum, and Wuppertal, generating open datasets that support digital public health research, urban planning, citizen science, and education.

Background

Conventional air quality monitoring infrastructures provide accurate measurements but are expensive and sparse, limiting their usefulness for fine-grained exposure assessment and urban health analysis. Advances in IoT-based low-cost sensorsenable denser monitoring networks but face persistent challenges related to calibration, sensor drift, and long-term data reliability. Within the Leibniz Digital Public Health Campus, AIR4LIFE builds on ongoing research in personal air pollution exposure monitoring and complements the SalusTransform initiative on health-promoting and equitable urban development. The project addresses the need for validated, reusable, and openly accessible air quality monitoring frameworks that can support evidence-based public health decision-making.

Research questions

• How can a low-cost, open-source IoT-based air quality monitoring framework be designed to ensure usability, reproducibility, and scalability?

• What are the capabilities and limitations of low-cost air quality sensors for reliable outdoor monitoring in real-world urban environments?

• How can AI-based methods improve sensor calibration, fault detection, and long-term data quality in distributed sensor networks?

• How can open, high-resolution air quality data support digital public health research, education, and urban planning?

Methods

AIR4LIFE develops a modular air quality sensor box based on the MoleNet IoT platform, integrating sensors for PM2.5, NO₂, CO, and VOCs. Sensor nodes are deployed in outdoor urban environments across Bremen, Bochum, and Wuppertal and validated through co-location with reference-grade monitoring stations. To enhance data reliability, the project integrates the AssureSense AI framework for anomaly detection, sensor fault identification, and data quality assurance. On-device sensor fusion, real-time calibration, and federated learning techniques are explored to improve robustness and adaptability across diverse urban contexts. All hardware

designs, software components, and datasets are released as open-source resources in accordance with FAIR data principles.

Expected outcomes

• A validated, low-cost, open-source IoT and AI-based air quality monitoring framework for urban environments

• An open, annotated urban air quality dataset with integrated quality indicators

• Publicly available hardware designs, software, and technical documentation

• Empirical evidence on the suitability of low-cost sensor networks for digital public health applications

• A scalable foundation for follow-up funding and larger interdisciplinary research initiatives

Project partners

• University of Bremen, Sustainable Communication Networks (ComNets) Group

• Wuppertal Institute, within the SalusTransform project

Project team

Prof. Dr. Anna Forster anna.foerster@uni-bremen.de University of Bremen

Prof. Dr. Gabriele Bolte gabriele.bolte@uni-bremen.de University of Bremen

Dr. Gibson Kimutai gkimutai@uni-bremen.de University of Bremen