SAFETY LEVEL RESEARCH AND RISK ASSESSMENT OF PERSONAL CARE WORKS

This study investigates the safety levels and risk assessment of personal care robots, focusing on their interaction with users and potential hazards. The increasing use of autonomous robots in personal care, particularly for elderly individuals, presents significant safety concerns. Although these robots are designed to assist users in daily activities, they pose risks such as mechanical injuries, electrical malfunctions, and privacy violations. Additionally, ethical concerns regarding user dependency and psychological effects are becoming increasingly relevant. The research categorizes personal care robots into three groups: mobile servant robots, transportation robots, and physical assistant robots. A comprehensive analysis of accidents involving these robots has identified the most common hazards, including sensor malfunctions, unexpected movements, and collisions. Additionally, the study evaluates existing risk assessment methodologies, such as Model-Based Safety Engineering (MBSE), Model-Based Safety Analysis (MBSA), and fuzzy inference systems based on ISO 13482 safety standards. A systematic risk assessment method was developed, incorporating checklists, surveys, and a risk matrix to quantify hazard levels. The evaluation of three specific robots—Lio (a mobile assistant with a robotic arm), iARM (a robotic arm mounted on a wheelchair), and KAIST (a wearable exoskeleton robot)—demonstrated varying levels of risk, with significant dangers associated with mechanical force, power failures, and unintentional shutdowns. Special attention was given to risks associated with prolonged human-robot interactions, including potential psychological stress and ethical dilemmas regarding autonomy and decision-making in caregiving environments. Findings indicate that while some risks can be mitigated through design improvements, others require enhanced user awareness and regulatory measures. Key recommendations include improving sensor reliability, implementing emergency stop mechanisms, and enhancing user training programs. Additionally, manufacturers should incorporate advanced human-detection algorithms and collision avoidance technologies to minimize physical harm. Furthermore, ethical considerations, including the psychological impact of robot interactions and data security, are discussed in-depth, highlighting the need for strict guidelines in robot deployment. In conclusion, this study underscores the necessity of integrating robust safety measures into personal care robotics to prevent accidents and ensure user well-being. Future research should focus on refining risk assessment models, improving human-robot interaction protocols, and advancing safety regulations for personal care robots. The continuous evaluation and adaptation of safety protocols will be crucial in addressing emerging risks and ensuring the widespread acceptance of these robots in healthcare and personal assistance sectors.