nctional electrical stimulation (FES) can used
for standing and stepping in complete spinal cord injury (SCI)
patients. A proposed method to enhance the performance of
FES gait is through the use of feedback control to modify the
muscle stimulation on a cycle-to-cycle or next-step basis. An
essential component of such a controller is the knowledge of
the timing of the gait cycle and gait events. The innovative
feature in this work is the use of a multi-level gait event
detector. The gait events are detected from insole force
measurements in real time. The lower level of the detector,
which is based upon previous classification work, estimates
the phase of gait at each time sample. The upper level of the
detector uses recently developed supervisory rules to detect the
transitions between the phases, i.e. the gait events. The gait
events of paraplegic patients walking with FES are detected in
real time. The gait events of most interest, heel strike and toe
offare detected within +9% of the gait cycle duration.