Neuroprostheses enabling patients to ascend and descend stairs can provide real functional gain. A novel finite state control scheme is presented and validated experimentally. The motion tasks Ascent and Descent are specified by sequences of motion phases for each leg separately. A supervisory controller coordinates motion phases and synchronizes left and right leg motion by Control Interpreted Petri Nets (CIPN). To switch between motion phases, sensor-based events and patient-operated manual switch states are supervised. This increases the patients' control of the movements and relieves them of automatizable tasks. Experiments with two complete thoracic spinal cord injured (SCI) patients were carried out. Transcutaneous stimulation of eight channels is applied by the experimental neuroprosthesis WALK! Both patients were able to repeatedly ascend and descend platforms of 12 cm and 16.4 cm in height, one patient also was able to ascend a staircase (17 cm step height). Feasibility of the approach has been successfully demonstrated. Future work will concentrate on improved foot clearance during Ascent and closed loop controlled muscle activation in selected motion phases of both Ascent and Descent. The approach is also suited for implanted lower extremity neuroprostheses.
Download full-text PDF |
Source |
---|---|
http://dx.doi.org/10.1046/j.1525-1403.2003.03007.x | DOI Listing |
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!