Decision Support System with AI-based Gait Estimation as Aid for Neurodegenerative Disease Patients
Decision Support System with AI-based Gait Estimation as Aid for Neurodegenerative Disease Patients

Author(s): Arun-Fabian Panaite, Monica Leba
Subject(s): Social Sciences
Published by: Udruženje ekonomista i menadžera Balkana
Keywords: IMU; Machine learning; Gait tracking; AI; Sensor fusion; Data fusion
Summary/Abstract: AI-based uncertainty handling can be applied to multimodal data fusion for IMU (Inertial Measurement Units) sensor-based gait motion capture in tracking gait differences in patients with Alzheimer’s disease or other medical conditions. The challenge is represented by monitoring and analyzing gait patterns in patients with Alzheimer’s disease to detect changes over time and assess disease, progression, or treatment effectiveness. Machine learning models are used to enhance the accuracy of gait analysis systems, making them valuable tools in healthcare for diagnosis and rehabilitation. Thus, IMUs have evolved with multi-sensor systems, sensor fusion, and machine learning for precise gait analysis, finding applications in clinical and consumer settings. AI-based gait motion capture has advanced through deep learning and video-based methods, enabling non-invasive, markerless analysis for individual identification, and enhancing healthcare diagnostics and rehabilitation. Recurrent neural networks (RNNs) or long short-term memory networks (LSTMs), are developed and trained using historical gait data from patients with Alzheimer’s disease that also include the uncertainty estimates as input features to the models. AI-based uncertainty handling integrated into gait motion capture and analysis allows continuous monitoring of gait differences in patients with Alzheimer’s disease and other medical conditions.

• Details
• Contents