GAIT

Gait disorders are among the most common and consequential symptoms of neurological conditions such as stroke and Parkinson's disease. They significantly impair mobility and quality of life and are a major cause of falls, with far-reaching health and economic consequences for both patients and the healthcare system.

Robotics Meets Artificial Intelligence

Robotic assistance systems have already fundamentally transformed neurological rehabilitation in recent years. They enable safe, intensive gait training, including practising stair climbing, standing up and sitting down, as well as fall prevention exercises. These systems continuously collect a wide range of valuable movement data. However, much of this data remains largely unused, as therapy planning is primarily based on therapists' experience and clinical guidelines. This is precisely where the GAIT project comes in.

The GAIT System: Personalised Therapy in Real Time

As part of the project, Fraunhofer IMTE and its partners are developing an innovative, data-driven decision support system for robot-assisted gait rehabilitation. The system uses signals from existing rehabilitation robots, with the FLOAT system from Reha-Stim Medtec GmbH serving as an example, to automatically determine clinically relevant movement parameters in real time. The GAIT system consists of four interconnected components:

• GAIT Assessment: A standardised, robot-assisted initial assessment identifies individual gait impairments, supported by virtual reality environments and targeted balance perturbations.
• GAIT Parcours: Based on the identified impairments, personalised therapy programmes are recommended and continuously adapted.
• GAIT Feedback: An integrated biofeedback system supports targeted therapy optimisation and addresses impairments in perception, coordination and sensory function.
• GAIT Control: The system actively controls the rehabilitation robot during therapy, for example by automatically adjusting body-weight support or perturbation intensity.

Our Contribution: Algorithms at the Intersection of Knowledge and Data

Fraunhofer IMTE's scientific focus is on developing algorithms to derive patient-specific movement parameters from robotic signals. We combine knowledge-based and data-driven approaches to achieve high precision and real-time performance while ensuring that the results remain transparent and interpretable for therapists.

Broad Applicability Through Modular Design

The GAIT system is designed to be modular and manufacturer-independent, allowing flexible integration into existing robotic platforms. The aim is to make personalised, data-driven gait therapy accessible to the widest possible patient population. The project initially focuses on ischaemic stroke and Parkinson's disease, two conditions for which the use of robot-assisted rehabilitation is increasing significantly and extensive data on pathological gait patterns are available. The project will conclude with a clinical feasibility study to validate the demonstrator, representing an important step towards medical device certification.