Bosch Sensortec launches 4-in-1 gas sensor powered by artificial intelligence

     Whether at home, in the office or outdoors, users want to be sure that the air around them is clean and safe to breathe - COVID-19 has demonstrated the importance of air quality. The ubiquitous particles, gases and even airborne viruses in the modern environment, and poor air quality are having an increasingly negative impact on our health.

     To meet this growing demand, Bosch Sensortec has released the BME688, the first MEMS air quality sensor that combines gas, humidity, temperature and air pressure sensing with innovative artificial intelligence capabilities, creating the world's most compact 4-in-1 air quality measurement solution. This standalone device enables customers to reduce total cost of ownership, shorten development time and simplify design.

     The new BME688 is ideal for innovative custom applications, such as detecting food spoilage by measuring current gases, tracking changes in temperature and humidity, or detecting forest fires in time. The sensor's artificial intelligence capabilities and Bosch Sensortec's new BME AI-Studio software tools allow customers to quickly develop customized solutions directly for their specific use cases. To further shorten time to market, Bosch Sensortec also offers Adafruit compatible development kits.

     Based on Bosch Sensortec's proven BME680 platform, the BME688's built-in updated gas sensor enables a wider measurement range and features advanced artificial intelligence capabilities. This gas sensor can now detect a variety of gases, including volatile organic compounds (VOCs), volatile sulfur compounds (VSCs), and other types of gases such as carbon monoxide and hydrogen, with a detection range of parts per billion (ppb).

     "The BME688 combines reliable, high-precision sensors with extensive gas detection and innovative AI capabilities." "This enables our customers to rapidly develop next-generation applications aimed at improving health, lifestyle and sustainability," said Dr. Stefan Finkbeiner, CEO of Bosch Sensortec.

     For example, a customer may want to develop a sensor-based product to detect spoiled food, which can be indicated by VSCs produced by bacteria in the food. It is also possible to detect bad breath or body odor based on VSCs produced by bacteria.

     The best approach is to collect indicative real-life data directly in the field. For example, by sampling gases near fresh and rotten food, different combination models of VSCs present in a given air sample are built. The BME688 detects VSCs very accurately by default, while the BME AI-Studio is optimized for other gas mixtures and applications.

     By sampling the gas in the field rather than in the laboratory, the new detection equipment uses a derivative algorithm that is more reliable in assessing the actual situation. Together with gas data, the BME688 measures humidity, air pressure, and temperature at the same time and serves as supplemental data input for complete AI modeling.

     In our example application, the customer classifies this large amount of data and then applies it to AI model development in BME AIStudio, that is, training BME688's ability to recognize obvious signs of bacterial growth on food. After the sensor training is completed, the final AI code is run on the system microcontroller of the customer's end product. This AI code is generally lightweight and can be easily run on microcontrollers currently used to handle system control and administration tasks.

     The BME688 has been developed for mobile and connected applications where sensor size and low power factors are highly demanding. The sensor is packed in a compact package, measuring just 3.0 x 3.0 x 0.9 mm3. Depending on the desired data rate and functionality, current consumption can be configured between 2.1 µA and 11 mA and optimized with BME AI-Studio.