As a core component of China's sports intangible cultural heritage, traditional martial arts face critical challenges in technical transmission, including movement distortion, non-standardized instruction, and cultural semantic loss. Existing motion capture technologies struggle to balance high-dynamic movement precision with cultural context interpretation (e.g., Tai Chi’s waist-hip rotation reflecting Yin-Yang philosophy). This study aims to design a multi-modal motion capture analysis system (MCAS) that integrates optical-inertial-force sensing architectures, culturally weighted algorithms, and real-time feedback mechanisms to resolve occlusion-induced drift errors and preserve both bio-mechanical accuracy and cultural semantics, thereby providing a technical pathway for digital preservation and transmission of martial arts. The study recruited 10 Tai Chi masters and 120 novice learners for three experimental phases in a laboratory setting: Technical validation: Simulated occlusion scenarios compared positioning errors between hybrid (optical+inertial+force) and optical-only systems; Cultural fidelity assessment: A double-blind experimental design evaluated the cultural accuracy of "Single Whip" movements under three conditions (traditional teaching, standard motion capture, and MCAS) via expert panel scoring; Pedagogical intervention: Learners were randomly assigned to MCAS or traditional groups for a 6-month longitudinal study, tracking skill acquisition speed, dropout rates, and cultural cognition changes. Data collection included kinematic parameters (200 Hz), ground reaction forces, and expert rating scalesanalyzed sic using mixed-effects models and thematic analysis for quantitative-qualitative integration. Technical performance: The hybrid system achieved a maximum positioning error of 2.3 mm under occlusion (vs. 8.7 mm for optical-only systems); Cultural preservation: The MCAS group attained 94.3% cultural accuracy (vs. 72.1% in standard and 65.4% in traditional groups), with 89% alignment of waist-hip rotation angles with Yin-Yang philosophy (vs. 54% in standard groups); Educational efficacy: The MCAS group showed 37% faster skill acquisition (p < 0.01), an 8% dropout rate (vs. 23% in traditional groups), and 41% higher cultural cognition scores (p < 0.001). This study pioneers the quantitative integration of cultural semantics into motion analysis algorithms, demonstrating that multi-modal fusion architectures overcome technical limitations of conventional motion capture. Compared to prior work, MCAS enhances data precision while addressing the critical issue of "formal resemblance without spiritual fidelity" through culturally weighted mechanisms. Limitations include reliance on expert-defined cultural weights, necessitating future AI-driven dynamic optimization. Practically, MCAS provides a replicable digital framework for safeguarding martial arts and other kinetic cultural heritage forms, with adaptable applications in dance, opera, and related domains.
Haoxiang Xie (Thu,) studied this question.