Biometric In-Cabin UX: Predictive Passenger Well-being via Multimodal Affective AI

Abstract

Car manufacturers are encountering increasing difficulties in comprehending relevant occupant experiences with regard to advancing their vehicles towards being autonomous mobile environments. Conventional feedback systems involving surveys and post-trip questioning provide only hindsight insights into an individual’s condition and neglect the dynamic momentary changes experienced during travel. A novel system is necessary to provide an ongoing biometric assessment and adaptive systems of intervention. Contactless sensors such as near-infrared cameras and thermal imaging systems measure biometric responses such as pupil reaction, facial expressions, and temperature changes without needing direct physical interface with car occupants. Machine learning algorithms interpret outputs from both biological responses as well as real-time driving factors such as acceleration and cornering forces to make predictive modifications of occupants' environment on an ongoing basis. Detecting predictive dissatisfaction indicators, such as heightened cognitive discomfort and initial stages of motion sickness, before a traveler's self-awareness of discomfort is necessary, is another function of this novel system that requires automatic adjustments of ambient lighting patterns, aural surroundings, and visual output levels to car occupants on an automatic basis. This novel system helps overcome trust issues on systems of autonomous driving of cars by ensuring validation of car occupants' status on an objective basis rather than a subjective basis of trust assumptions. Building a cognitive interactive interface for control systems of a car and biologically interactive responses of an individual leads to designs of more optimized car environments that provide for sustained comfort and trust of car drivers during their travel experience through novel systems of autonomous car driving environments.