Toward an AIoT-Based Body Sensor Network for Learning Tai-Chi Chun

Abstract

Tai Chi, an ancient Chinese martial art, is known for its benefits in promoting physical and mental well-being, such as slowing aging, preventing diseases, and alleviating stress. However, learning Tai Chi can be challenging, especially for beginners without an instructor’s guidance. Mastery requires understanding complex principles like body coordination, weight shifting, and force exertion, which are essential for proper technique. Traditional training relies heavily on experienced instructors, making it difficult for individuals to practice independently and progress effectively. This study addresses the challenge faced by beginners who struggle to improve without continuous feedback, particularly in the absence of an instructor. Here, we propose an artificial intelligence (AI)-based system to guide and correct students’ movements using wearable sensors, including inertial measurement units (IMUs) and smart insoles, along with smartphone cameras. The system collects motion data from both instructors and students, compares their movements, and provides immediate feedback on discrepancies. Unlike other Tai-Chi training systems that focus mainly on posture correction, our system emphasizes force exertion—a key component of Tai Chi. By analyzing movements and weight distribution, the system helps students identify errors in technique, improving their force generation and body coordination. This feedback mechanism enables independent practice, making Tai-Chi training more accessible and efficient.