An overview of the feasibility of permanent, real-time, unobtrusive stress measurement with current wearables

Abstract

Negative consequences of stress are a pervasive problem in our modern society. Recent developments in wearable lifestyle hardware have led to unobtrusive, sensor-packed, always-on devices that might finally be able to continuously monitor biosignals to detect, determine or even prevent stress or some of its negative outcomes. In this work, we give a concise overview of a majority of biosignals that are in some way relevant for stress classification and outline state-of-the-art machine learning algorithms for this task. Additionally, we provide a list of all recent wearables including an evaluation of their feasibility to implement such algorithms as well as directions to look for an assessment of the accuracy and validity of their recorded data with respect to stress tracking.

Publication
ACM International Conference Proceeding Series

Related