Machine Learning Enhancing a Compact Wearable Device for Stepping Management

Walking is a complex process requiring various musculoskeletal muscles' synchronized actions. Abnormal gait increases the risk of joint deformities, and foot pronation or supination is a common health issue. Wearable technologies for health monitoring have gained significant momentum. This study used an IMU to collect components of foot location (x, y, z) for detecting gait deformities such as pronation and supination. A machine learning classifier was developed to categorize different types of gaits: pronation, supination, and normal. The classification process was conducted in two phases. In Phase 1, the classes were pronation, supination, and normal. In Phase 2, the classes were severe pronation, pronation, normal, supination, and severe supination. The highest accuracy achieved in Phase 1 was 97%, while Phase 2 reached an accuracy of 99.2%. The novelty of this work lies in the sensor location, which is more convenient for the user than the current in-sole sensors. The proposed solution effectively corrects foot posture and is particularly efficient for cohorts like the elderly and children.