Bernstein Focus on Neuro Technology
Grant agreement no.: 01GQ0810
Grant Scheme: Federal Ministry of Education and Research (BMBF) - National Bernstein Network Computational Neuroscience
Project full title: Bernstein Focus Neurotechnology - Neurobionische Kontrollsysteme
Project acronym: BFNT
Duration of the project: 60 months
Start date of the project: 01.10.2008 (in Neurorehabilitation Engineering from 01.11.2010)
Project successfully completed
Within the Bernstein Focus Neurotechnology (BFNT) Goettingen, the Department of Neurorehabilitation Engineering aims at the development of closed-loop prosthetics systems controlled by surface EMG signals. Such novel systems should be intuitive to use, allow dexterous control, provide an artificial sensory feedback to the subjects, and be robust enough for being commercially produced and clinically used.
The four main areas of development on which the project will focus to improve the current prosthetics systems controlled by EMG signals are the following:
- Adaptation. Algorithms for the control of prosthetic devices should adapt to the changed signal features during the use of the prosthesis. Examples of factors that determine variability in signal features are muscle fatigue, displacement of the electrodes, and arm posture;
- Closed-loop control. The sensory information is necessary for dexterous control, as it is evidenced by de-afferented patients, who often cannot perform any movement despite the intact efferent pathway. Active prostheses should thus include artificial sensory information delivered to the user. Such information is currently missing in commercial and clinical devices;
- Simultaneous and proportional control. The academic state-of-the-art in myoelectric signal classification allows discrimination between >10 movements with accuracies >95%. However, such pattern recognition scheme is inherently sequential (one class at a time) and on/off (without control on speed or force). This is in contrast with the natural way of limb control which is simultaneous and proportional over multiple degrees of freedom;
- Multi-modality. The EMG signal may be used together with other signals for improving the dexterity in prosthesis control. For example, video-cameras mounted on the prosthesis may allow the recognition of the object shape and size for optimizing the grasping without direct control of the user (who would provide higher-level control signals)
Within the BFNT Goettingen, the Department of Neurorehabilitation Engineering is working toward solving the challenges in the above 4 key areas to provide better control of artificial limbs than it is currently possible.