Textile emitter for AI-enhanced human–machine interaction

Abstract Wearable textile interfaces hold significant promise for advanced human–machine interaction, yet realizing wireless, battery-free on-body signal generation without embedded circuitry in textiles remains a formidable challenge. Here, we present a textile emitter that enables precise interaction by directly transducing user-induced mechanical deformation into time-coded infrared (IR) thermal signals. Constructed from strain-responsive IR-emitting fibers, the textile emitter simultaneously functions as a programmable signal generator without embedded electronic circuits or power sources within the textile. Seamlessly integrated into everyday garments, it ensures stable performance while maintaining comfort and wearability. Inspired by Morse code, user motions are encoded as temporally structured IR signals and wirelessly decoded by machine learning algorithms, achieving up to 99.6% recognition accuracy across diverse interaction commands. This AI-enhanced textile emitter establishes a scalable and robust foundation for future wearable systems in information communication, robotic control, and digital interaction.