Motor relearning after stroke is a lengthy process and aims to help patients recover voluntary motor movement capability so that they can return to an independent lifestyle. Numerous therapy approaches exist to ensure suitable conditions for motor recovery after stroke. Robot-based therapy has become a widely used therapeutic approach in stroke rehabilitation, which supports the patient's motor learning through highly intensive movement training using a rehabilitation device. Haptic human-robot interaction is a key factor that influences motor learning in robot-based therapy because it defines how robots actually support patients. It is important, that robotic assistance is adapted to the individual patients’ abilities, thereby ensuring high efficiency of therapy. There is a lack of evidence to assess which algorithm is preferred for motor recovery and how to define the best control strategy for a therapeutic device. I hypothesised that natural and learning-effective human-machine interaction can be achieved by programming the robot’s control so that it emulates how a physiotherapist adaptively supports the patients’ limb movement during stroke rehabilitation. Therefore, one of the aims of this doctoral thesis is to examine the possible benefits and/or disadvantages of physical human-human interaction in general and therapist-patient interaction in particular for motor learning and acquire the first findings of the haptic behaviour of trained therapists. Additionally, the influence of haptic human-human-interaction on human perception and performance was the focus of this project. To investigate these objectives, a haptic interface Bi-Manu-Interact was designed and implemented as well as three experiments that investigated perception, motor learning and recovery during the stroke rehabilitation were conducted. The experimental investigations showed that humans prefer trajectory guidance that is widely used in rehabilitation as a partner for the joint achievement of a motor task. Surprisingly, the subjects' least preference for interaction was a human partner that they characterised as disturbing and unpredictable. Individuals could also distinguish guidance from human and human-like robot partner, which were both more compliant reactive agents. However, further examination of control algorithms showed that haptic guidance may impede short- and long-term motor learning and the transfer of the acquired skills. In contrast, interaction with a human established itself as the most beneficial of all agents for improvement in smoothness and accuracy. A long-term clinical study supported these findings since three-weeks intervention including jointly training with occupational therapists over a haptic interface enhanced patients' smoothness and accuracy of movements. Overall, the results of this doctoral thesis support the hypothesis that haptic human-human interaction positively influences the motor learning of healthy individuals and stroke patients. Therefore, the modelling of such interaction could be potentially beneficial for control applications in rehabilitation robots. However, this should be in-depth investigated in further research.
Control for robot-assisted neurorehabilitation based on human interactive behaviour
Regelungsalgorithmen für robotergestützte Neurorehabilitation basierend auf menschlichem interaktivem Verhalten
2021
Miscellaneous
Electronic Resource
English
DDC: | 629 |
Customized Robot Aided Gait Training for NeuroRehabilitation
Springer Verlag | 2013
|Monitoring visual attention on a neurorehabilitation environment based on interactive video
BASE | 2013
|A Novel Robot Neurorehabilitation for Upper Limb Motion
IEEE | 2005
|