Full paper accepted in top CS conference - ACM Computer Human Interaction, CHI 2018

FAST AND FURIOUS: DETECTING STRESS WITH A CAR STEERING WHEEL

January 8, 2018

Abstract

Stress affects the lives of millions of people every day. In-situ sensing could enable just-in-time stress management interventions. We present the first work to detect stress using the movements of a car’s existing steering wheel. We extend prior work on PC peripherals and demonstrate that stress, expressed through muscle tension in the limbs, can be measured through the way we drive a car. We collected data in a driving simulator under controlled circumstances to vary the levels of induced stress, within subjects. We analyze angular displacement data to estimate coefficients related to muscle tension using an in- verse filtering technique. We prove that the damped frequency of a mass spring damper model representing the arm is significantly higher during stress. Stress can be detected with only a few turns during driving. We validate these measures against a known stressor and calibrate our sensor against known stress measurements.