A smart device to collect breath data in real-life situations
Intelligent Systems Lab, Sonoma State University, Rohnert Park, CA, USA
Research and development in breath-based diagnosis and monitoring devices for diabetes and other diseases have recently seen an exponential growth. This has created a need in the scientific communities for a sensor device that allows a large scale breath data collection from patients in real-life situations. This is particularly important as the advances in machine learning and artificial intelligence has enabled complex sensor pattern based recognitions and classifications. The sensor device reported in this invention has a small form-factor (handheld) that can take breath input, record sensor readings, and wirelessly send data to the cloud. The device is easy to use and can be given to the patients to conduct self-tests with breath at home. The device has three slots for chemical/volatile organic compounds sensors (which can be expanded as needed), sensor interface that can be programmed to match the sensor characteristics, and a program for automatic calibration of sensors. The device comes with three VOC sensors, which can be switched out based on the study need. The device can be easily programmed to package the data, format it, send it to the cloud for automatic logging on cloud databases. The microcontroller on the board can be reprogrammed with trained algorithms for real time classification of patient breath and make predictions. Thus, this device can be used for verification and further training of the algorithm, as well as implementing the trained algorithms and turning into a monitoring and/or diagnostic device without any hardware modifications.