Task-oriented and Purposeful Robot-Assisted Therapy

摘要

This chapter discussed the development of a task-oriented therapy robot focused on real ADL training and performance. The development of the software, HERALD, and the hardware platform and FES grasp glove has been discussed. We also presented training and trajectory planning models that have been implemented with the system along with defining the pros and cons of three possible models that can be used to accurately reflect natural movement of the wrist during ADL task. We presented three case studies that briefly examined how the system has repeatable performance and has the ability to train stroke survivors. A low functioning stroke survivor was successfully trained on the system using PTP ADL-like movement. The subject's kinematics, especially movement time and movement smoothness decreased reflecting motor impairment reduction and increased motor control. The ADL functioning was improved on tasks involving more shoulder and elbow function but not on task involving grasp. In the future, the FES glove will be fully integrated in the ADLER system to allow stroke clients on all levels of motor function and ADL ability to be trained in both reaching and grasping ADLs. As such we anticipate that a major impact will then be seen on both motor impairment and functional scales. We also presented via case study 2 different trajectory models for planning and assisting movement of the wrist using ADLER. We demonstrated that models that included curvature and had customized inputs can improve a subject's movement kinematics for an ADL tasks such as drinking and can affect their perception of the ease or difficulty of the movement. In the case study performed on the ADLER system, the subject reported a "more natural" feel when operating with the new model rather than the old model. This shows that the model appears to meet the goal of providing a more natural prediction of functional