Quantifying neuromechanical parameters during a robot assisted visuomotor tracking task

Quantifying neuromechanical parameters during a robot assisted visuomotor tracking task

Koenen, B.W.

Van der Helm, F.C.T. (mentor)
De Vlugt, E. (mentor)

Mechanical, Maritime and Materials Engineering

BioMechanical Engineering

BMD

2015-06-17

In the field of neurorehabilitation, robots are being used to provide assisting forces for therapeutic purposes, and with system identification & parameter estimation (SIPE) methods they are used to quantify neuromechanical parameters for diagnostic purposes. To better understand the beneficial effect of robotic assistance on the neuromechanical system during therapy, a tool is required that combines SIPE methods with robot assisted training tasks. This experimental study proposes a novel protocol that demonstrates the potential combination of quantifying neuromechanical parameters, using a quasi time-variant SIPE method, with a robot assisted visuomotor tracking task. A haptic wrist robot was used that simultaneously exerts assistive forces by means of a virtual spring, and force perturbations required for SIPE of the wrist joint. The tracking task was visually fed back through a gaming environment. The effect of robotic assistance on human effort and task performance was investigated and further clarified by the underlying neuromechanical behavior. Six healthy controls and one chronic stroke patient participated to test the protocol. The quasi time-variant SIPE method returned sufficient estimations with an average VAF of 86%. The robotic assistance resulted in an average increase of task performance but an average decrease of human effort. Findings of the neuromechanical parameters suggests that the neuromechanical system tended to behave more compliantly with robotic assistance. While the tracking task was presented as a position task, findings further suggest that higher magnitudes of robotic assistance might have elicited neuromechanical behavior more similar to force tasks. The findings are in line with the paradigm of shared haptic control which says that humans adapt their neuromechanical impedance to promote task performance when exchanging forces on the control interface of the human-robot system. It is concluded that the proposed protocol simultaneously yields methods for combined motor training and quantitative assessment of the wrist joint. Moreover, the novel quasi time-variant SIPE approach showed to be a promising tool to quantitatively assess stroke patients during intensive and frequent motor training tasks.

robot assistance
SIPE
neuromechanical parameters
motor training task
stroke

http://resolver.tudelft.nl/uuid:41206883-5c79-4749-a47a-ea43c843677a

2015-06-21

Student theses

master thesis

(c) 2015 Koenen, B.W.