At-home monitorization using artificial noses and multimodal sensors: a noninvasive approach to human routine certification.

Abstract

Latest advances in e-health highlight the importance of gathering accurate data from patients, to allow for a better diagnosis, and in turn provide more effective treatments. The remote communication with patients however makes it difficult to gather accurate information, as it is often biased or limited to what the patient can communicate to the doctor. We have developed a noninvasive integrated system that can monitor human routines more effectively in a noninvasive manner using multimodal sensors. The human body releases multiple substances into the air through its natural biological functions. Breathing, sweat, digestion, etc., all release chemical components (I.e., CO2, amines, methane) which we readily measure with specialized olfactory sensors, effectively “artificial noses”. Our device registers the variability in these air parameters associated with cognitive activity and complements this data with other multimodal sensors (presence sensors, luminosity, loudness, etc.). In the integrated system proposed the data is then transmitted in real-time into a server where it can be securely stored, even for years, and quickly accessed and analyzed. We have tested the use of the proposed noninvasive monitoring technology in university classrooms and in a primary school. Our results show that all routines in these environments are reflected in the sensor signals and that the artificial nose can be used to certify the corresponding cognitive activities. The noninvasive approach allows to obtain valuable data over long periods, and in the context of e-health this could allow a much better understanding of patient routines and the changes associated with health conditions. Also, the real time nature of this approach allows to implement “early warning” or notification strategies to quickly react upon changes in routines that signal a health problem. Artificial noses and associated information systems to store and analyze the data online and offline can be used to noninvasively monitor human activity and, in particular, cognitive activity. We argue that this technology can be directly applied to monitor the elderly, people at early stages of neurogenerative diseases, any progress in cognitive disease development, and to ensure that patients correctly follow treatment.