Diabetes Detection Using Gait Analysis and Machine Learning

Diabetes mellitus (DM) is a medical condition affecting multiple organs and systems due to high blood glucose levels in the body. Diabetic neuropathy (DN) is a common consequence of DM, which causes sensitivity loss in the nerves of muscles. Another complication of DM is diabetic foot ulcer (DFU), which indicates the worsening of DN. Conventionally, diabetes is detected using invasive methods. However, invasive methods are not convenient and can cause pain to patients. Thus, an alternative method to detect diabetes is to use gait analysis. Gait sensors used in gait analysis can be categorised into kinematic and kinetic sensors. Kinematic sensors include accelerometers, shoe-type IMUs, ear-worn inertial sensors and motion capture systems, whereas kinetic sensors include force-measuring shoes, pressure sensors, EMG sensors and force plates. Although gait sensors provide many gait features such as acceleration, plantar pressure, EMG and ground force reaction, manual gait data analysis can be challenging. Therefore, machine learning (ML) and deep learning (DL) algorithms are used to analyse gait data. Most papers reviewed have used ML models in detecting diabetes using gait data. However, DL models are not commonly applied in diabetes detection. Thus, future studies should explore and implement DL models to evaluate their performance in diabetes detection.