TinyML Models for a Low-cost Air Quality Monitoring Device


Low-cost air quality monitoring devices can provide high-density spatiotemporal pollution data, thus offering a better opportunity to apply machine learning (ML). Low-cost sensor nodes usually utilize microcontrollers as the main processors, and tinyML brings ML models to these resource-constrained devices. In this letter, we report the development of a low-cost air quality monitoring device with embedded tinyML models. We deploy two tinyML models on a single microcontroller and perform two tasks: predicting air quality and power parameters (using model predictor) and imputing missing features (using model imputer). The proposed model predictor can estimate parameters with a coefficient of determination above 0.70, and the model imputer effectively estimates the testing data when missing rates are below 80%. By performing the posttraining quantization technique, we can further reduce the model size but slightly degrade the accuracies.

IEEE Sensors Letters, vol. 7, no. 11
I Nyoman Kusuma Wardana
I Nyoman Kusuma Wardana
Warwick PhD Alumnus

My research interests include machine learning for environmental systems.

Suhaib A. Fahmy
Suhaib A. Fahmy
Associate Professor of Computer Science

Suhaib is Principal Investigator of the Accelerated Connected Computing Lab (ACCL) at KAUST. His research explores hardware acceleration of complex algorithms and the integration of these accelerators within wider computing infrastructure.