Emotion-Driven IoT Feedback Loop for Caregiver Training

Abstract

Emotional sensitivity is one of the pillars of effective therapy of autism, but caregivers do not have real-time instructions on how to interpret affective expression of children. This paper proposes an Emotion-Driven IoT Feedback Loop (EDIFL) that is focused on feeding back data-driven feedback immediately to caregivers during behavioral interventions. The system includes multimodal IoT sensors, edge level emotion recognition and cloud-based adaptive feedback generation. The framework is made more responsive and accountable in autism care, through integrating machine learning, human-centered AI, and NIST AI RMF governance. Findings indicate an increase of 47% in accuracy of caregiver response and decrease of 38% in delayed reactions when compared to the old method of training sessions. The suggested EDIFL model is a great advancement in terms of intelligent, compassionate and morally responsible caregiving assistance.