Research on a Multimodal Intelligent Dressing System and Digital Age-Friendly Design Based on Natural Behavior Theory

INTRODUCTION:With global population aging, everyday dressing decisions for older adults have evolved from simple garment choices into complex tasks that must simultaneously satisfy health, safety, social etiquette, and subjective comfort constraints. Experience-based and intuitive dressing strategies struggle to cope with rapidly changing weather, fluctuating health conditions, and frequent switches between daily events, often leading to misalignment between clothing choices and healthy aging goals. OBJECTIVES:This study proposes an intelligent dressing recommendation system for older adults based on natural behavior theory. The system adopts a WEHT (Wearable–Event–Health–Thermal environment) situational modeling framework and multimodal data fusion to provide safe, comfortable, socially appropriate, and easy-to-understand dressing recommendations within a digital age-friendly design context. METHODS:The system constructs an elderly-oriented clothing knowledge graph and functional matrix, and fuses multimodal inputs including weather indices, individual health constraints, and calendar events. A cross-modal attention mechanism with adaptive weights is introduced to capture direct and indirect couplings among clothing, events, health status, and thermal environment. Under a predefined hierarchical decision rule that prioritizes health and safety, the system employs multi-objective optimization and fuzzy rules to generate explainable and executable dressing plans. Recommendations are presented through natural interactions combining large graphical icons, step-by-step guidance, and voice prompts. A user study with 100 participants aged ≥65 years was conducted, including usability testing and controlled comparison with traditional self-decision dressing. Outcome measures covered health risk avoidance, scenario adaptation accuracy, decision time, and subjective satisfaction. RESULTS:Compared with the traditional experience-based dressing strategy, the proposed system significantly improved health risk avoidance and scenario adaptation accuracy, shortened dressing decision time, and increased subjective satisfaction across multiple scenarios. Older participants were able to understand and follow the system’s recommendations with relatively low cognitive load, benefiting especially in complex or health-sensitive situations. CONCLUSION:This study integrates natural behavior theory with multimodal intelligent algorithms under a digital age-friendly design paradigm, and proposes a “behavior–context–environment coupling” human–AI co-decision framework for intelligent dressing. The results demonstrate that the WEHT-based multimodal dressing support system can effectively enhance safety, comfort, and contextual appropriateness of clothing decisions for older adults, while providing a theoretically grounded and practically feasible pathway for intelligent life-assistance systems in healthy aging.