中国科学院自动化研究所机构知识库

Active rehabilitation training based on recognition
of motion intention can effectively improve the patient’s engagement in rehabilitation training, and thus improve the training
effect. In this paper, a speed adjustment strategy for active
training based on multi-joint damping control is proposed for
bicycle training. The active force applied by patient can be
calculated by the dynamic model of human-robot system, and
then converted into the tangential force at the pedal of robot
along the forward direction of bicycle. The tangential force is
converted into the adjustment term of joint angular speed by
damping control to dynamically adjust the riding speed. When
the tangential force is larger than the threshold value, the pedal
will deviate from the reference circular trajectory. Therefor, a
speed vector pointing to the circular center is added to pull
the end-effector back to the reference trajectory. Moreover, a
fuzzy impedance parameter regulator is designed to adjust the
training intensity, by which the impedance parameters can be
regulated according to the magnitude of the patient’s active force
and the deviation from the reference trajectory. Finally, in order
to increase the patient’s engagement, Unity3D software is used to
design the virtual scene of cycling on the road. The experimental
results show that the active compliant rehabilitation training can
be realized by the proposed method.