Electronic Nose Based on AI-capable Sensor Module for Beverages Identification

In this paper, we present an electronic nose application case study based on an affordable Bosch BME688 AI (artificial intelligence) capable sensor module set (and development kit) to distinguish and identify different soft drinks and beverages. During the work, the device was prepared for TinyML and AI-based identification. The module contains eight pieces of BME688 sensors, and it is compatible with the Bosch BME AI Studio software. Based on their unique odor profiles, we created and tested various ML (machine learning) algorithms to distinguish between fruit palinka (fruit spirit) types, different beers, and ethanol. The case study aims to determine whether the module can distinguish between beverages. For this, we applied passive delivery. It was found that the differentiation between air and different samples (cola, lager beer) is possible. Still, ethanol and spirits with high alcohol content saturate the device, pointing to future dynamic sampling (active delivery).