AI-assisted personal training gear

Abstract

Proper posture and muscle engagement during exercise are crucial for maximizing training efficiency, minimizing unnecessary stress on ligaments and joints, and lowering injury risks. However, existing wearable technologies fall short in meeting the demands of high-intensity exercises. This study introduces a personal training gear that leverages synchronized and continuous electromyography (EMG) signals, inertial measurement unit (IMU) data, and imaging inputs to monitor body movement and muscle activity during exercise. These objective measurements offer a comprehensive, real-time overview of training performance, assisting personal trainers and athletes in optimizing training programs to enhance efficiency and minimize injury risks. The research focuses on optimizing sensor placement and interfaces to ensure accurate data collection, as well as designing patterns and selecting fabrics that enable freedom of movement and wearing comfort. Machine learning and neural network algorithms were developed to (1) detect eight specific exercises and their common pose deviations using video imaging, and (2) assess Muscle Fatigue and predict the estimated number of remaining repetitions with EMG signals. The system also recommends the optimal training load to maximize efficiency. By monitoring muscle utilization patterns and body motion, the gear identifies incorrect exercise poses and provides real-time warnings to the user. This AI-assisted personal training gear addresses critical technological gaps in current wearable systems, particularly in monitoring and enhancing strength training targeting specific muscle groups. Its advanced measurements and analytics offer a comprehensive evaluation of training performance in real-time, with applications spanning the healthcare and sports industries. Patents are filed in China, Hong Kong and the United States. The research outputs have been presented at the International Conference on Applied Human Factors and Ergonomics (2022 & 2023), TechConnect World Innovation Conference and Expo i n Washington, DC, exhibited in Paris, and featured in local media reports.