Motion tracking systems in rehabilitation: a comparative analysis of sensor technologies, algorithmic methods, and clinical relevance

Abstract

Motor impairments are a common consequence of neurological and orthopedic conditions and often require long-term monitoring and rehabilitation. Digital motion-tracking systems — including optical marker-based solutions, optical markerless systems based on red-green-blue cameras and depth sensors, wearable inertial measurement modules, and their hybrid combinations — enable quantitative assessment of body kinematics, detection of compensatory movement patterns, and objective tracking of patient recovery. This review systematises modern sensor technologies for motion tracking in rehabilitation, evaluates their metrological characteristics, algorithmic approaches for data processing and interpretation, and determines their clinical relevance and development pathways. A comparative analysis of four main classes of systems shows that optical marker-based technologies provide the highest spatial accuracy but remain limited by high cost and laboratory dependence; markerless optical systems demonstrate moderate accuracy yet offer promising opportunities for telerehabilitation; inertial measurement modules provide mobility and reproducibility but require drift-compensation algorithms; hybrid multisensor architectures combine the advantages of both optical and inertial approaches and achieve high accuracy even under dynamic conditions. The review also analyses sensor-fusion algorithms such as extended Kalman filters and Madgwick filters, methods for computing kinematic parameters, and machine-learning models for automated motion interpretation, including explainable artificial-intelligence frameworks that improve model transparency and support clinical decision-making. Clinical-validity assessment indicates that commonly used metrics such as range of motion, gait-phase characteristics, and angular parameters show high consistency, whereas more complex indicators may vary and require cautious interpretation. Key challenges include the lack of standardised measurement protocols, environmental influences, limited interpretability of algorithms, and difficulties integrating systems into clinical workflows. Promising future directions include the development of multimodal datasets, adaptive sensor-fusion strategies, transparent artificial-intelligence analytics, and scalable home-based rehabilitation solutions, highlighting the potential of integrated multisensor technologies to support personalised rehabilitation.