The way back to a normal life is very difficult for stroke or accident patients. Simple movement patterns like walking to the post box or picking up a coffee cup can turn into insurmountable obstacles. Early, regular rehabilitation is often the deciding factor in how well patients can re-accustom themselves to everyday life. Walking or arm therapy is very laborious and needs extreme exertion by both the therapist and the patient, which imposes time limits on it. Various rehabilitation robots offer helpful support. The “Lokomat” is one of these. With its help, movements can be performed more often and with different intensities.

Motivating patients even better

Rehabilitation robots are now undergoing further development. A project entitled “Multimodal Immersive Motion Rehabilitation with Interactive Cognitive Systems”, or MIMICS for short, has brought together leading researchers from Robotics, Presence Research and Neuro-rehabilitation. Project leader Robert Riener, Professor at the Institute of Robotics and Intelligent Systems of ETH Zurich and Professor at the Spinal Cord Injury Center of the University Hospital Balgrist, summarises the MIMICS objective: “The aim of the new method is to employ virtual reality in such a way that a stroke patient undergoing training receives maximum motivation, and the therapy exercises are adapted automatically and continuously to the patient’s ability and attentiveness.”

Up to now robots have been able to carry out various exercises with patients during rehabilitation. In future the plan is to observe a stroke patient’s attentiveness and the intensity of the exercises via sensors, thus enabling direct intervention in the progress of the exercises. Robert Riener explains that “The patient walks on a treadmill. In front of her is a projection screen on which she sees a path along which she must walk. The path shows virtual obstacles, for example a curb edge, which she must climb over. If our sensors show us that the patient’s attentiveness is decreasing or the exercises are too easy, we can change the exercise procedure automatically.”

Training with all the senses

The patient is stimulated through all her senses during the training. For example if she walks across a wooden floor, an appropriate noise is heard from the loudspeakers. If her attention slackens, the patient is shown a ball on the screen and must kick it into a goal. And if she walks too slowly, a person appears in the virtual world and takes the ball away from her. The new systems interact with the patient, thus adjusting continuously to their physical and mental state without overtaxing or failing to present sufficient challenge. This feedback mechanism enhances the quality and intensity of the training and thus the patient’s chances of rehabilitation.

The aim of the research project, which lasts three years, is to find out how these various pieces of information can be retrieved, assessed and fed back into the system as new parameters. It involves studying three parameters in particular: the patient’s physiological values, biomechanical data and behaviour. Among the physiological values, for example, the pulse rate is measured, and for biomechanics the exercise force exerted by the patient is evaluated. At the same time the system checks the direction in which the patient is looking and how attentively she is following the training. The technical challenge in all of this is to network all the data together by sensor merging, and to evaluate it in real time.

Project funded by the EU

The project receives 1.6 million euro of funding from the EU for three years via the “7th Framework Programme”. The entire project volume amounts to 2.2 million euro. About 120 projects applied for EU funding. Only ten percent have now been accepted. For Robert Riener the award is also proof of the project partners’ professionalism: “It is not easy to gain acceptance in a competitive environment as difficult as this. The long-term interdisciplinary research gave us good pre-conditions.” In addition to the project leadership by Robert Riener, Professor at the Institute of Robotics and Intelligent Systems of ETH Zurich and Professor at the Spinal Cord Injury Center of the University Hospital Balgrist, the following partners are taking part in the project: Professor Volker Dietz from the University Hospital Balgrist, Professor Marko Munih from the University of Ljubljana in Slovenia, Professor Mel Slater from the Universtat Politècnica de Catalunya in Spain, Dr. Friedemann Müller from the Neurological Clinic of Bad Aibling in Germany, and the Hocoma AG company in Volketswil as the industrial partner.