Metaverse‐Enabled Yoga Coach Avatar Using AI‐Enhanced Multimodal Insole Sensing System

Abstract Plantar biomechanical monitoring has emerged as an indispensable tool for health assessment and activity recognition. However, existing insole systems lack the capability to support multimodal sensing for sports monitoring due to limitations in materials and complexity of system design. Hydrogel, owing to its multimodal sensing capabilities and biocompatible properties, showcases great potential for advanced plantar monitoring during exercise. Here, the study proposes an artificial intelligence (AI)‐enhanced multimodal insole sensing system (AEIS) based on ionic hydrogel for immersive, real‐time posture correction and personalized yoga training guidance. The AEIS integrates a 32‐channel hydrogel‐based sensing array with a customized wireless circuit, enabling simultaneous monitoring of plantar pressure, temperature, and sweat. By leveraging hybrid AI algorithms, AEIS serves as a virtual yoga coach, achieving high accuracy in posture recognition (98.33%) and imbalance detection (90.06%). Furthermore, the developed approach based on random forest (RF) is trained on center‐of‐pressure (COP) stability data from yoga coach, enabling AEIS to analyze real‐time data of yoga practitioners and deliver personalized posture guidance. Meanwhile, the embedded haptic units provide real‐time haptic feedback in response to improper plantar pressure distribution. AEIS forms an interactive metaverse‐based yoga training platform, offering users an immersive, face‐to‐face‐like experience with a virtual yoga coach.