An Adaptive AI and Smart Textile Platform for Affective Human–Robot Touch Interaction

This interdisciplinary project presents the development and validation of adaptive, AI-powered plush robot, prototyped as pets. Our core innovation is a sensitive tactile perception system built on smart textiles, capable of distinguishing and interpreting a wide spectrum of affective human touch in real time. This is enabled by an embedded machine-learning algorithm, which achieved 75% accuracy in classifying distinct affective gestures during systematic testing.

Empirical research has validated the prototype's effectiveness in structured interventions for children with Autism Spectrum Disorder, where it fosters secure tactile engagement and communication. A key strength of this work lies in the demonstrated adaptability of both the smart textile-based hardware and the core AI algorithm. This foundation enables significant customization, including the development of new physical embodiments beyond the current animal companions, the recognition of context-specific gestures, and tailoring for diverse user groups across various therapeutic and support contexts.

We envision broad potential for this technology as a complementary, human-centric tool within therapeutic ecosystems. Potential applications extend to areas such as senior care, services for individuals with disabilities, crisis intervention, and psychosocial support, among others. We are actively seeking collaborative partnerships to co-create and pilot tailored solutions, leveraging our proven sensory and AI foundation to explore new applications and address the needs of different user groups.