IEEE Robotics & Automation Magazine - December 2017 - 63

early trials but developed appropriate responses later in the
process. After this preliminary study, we determined that individuals presenting traumatic injury above the 12th thoracic
vertebra (T12) level were the most eligible candidates to take
part in the ensuing preparation protocol.
Based on the initial results, our pilot (a 38-year-old male,
with American Spinal Injury Association Impairment Scale
designation A, level T9, who had been injured three years
earlier) was selected from this group. We conducted a series
of exams to confirm his tolerance for the preparation
protocol, which indicated that he had no history of bone
fractures, 78% of normal bone density preserved (considering gender, age, and height), a complete range of joint
motion, an absence of muscle contractures, normal cardiovascular response to stress, healthy skin and no sores, adequate metabolic condition, and an absence of autonomic
disrreflexia or postural hypotension. In addition, the pilot
demonstrated a high level of motivation, particularly after
feeling his lower limbs cycle again.
After his screening examination and clinical evaluation,
the pilot's FES training began. At the beginning of his prepreparation phase, not more than 10 min of satisfactory muscular contraction were possible due to the early onset of
FES-induced fatigue. These initial sessions were necessary to
create muscle adaptations to the electrical stimuli, in particular for the conversion of type I to type II muscle fibers. Also in
this prepreparation phase, grade two contractions were
obtained according to the Medical Research Council's scale
for muscle strength (i.e., muscle contraction is visible but it
does not provide gravity support). During the fourth session,
grade three contractions were reached, indicating minimal
support against gravity.
In the first preparation phase, which lasted 18 weeks, isometric exercise sessions were increased to five a week, and our
pilot adapted to new modalities of exercise, including dynamic contractions and functional tasks, such as sit-to-stand,
walking, and, finally, cycling.
Regular FES cycling training, the second preparation
phase, started 12 weeks before the competition. This phase
featured three weekly sessions of first stationary and then
mobile FES cycling. With regard to the stimulation parameters, we first kept the frequency fixed at 50 Hz but then
decreased it to 30 Hz after the initial sessions in an attempt
to delay the onset of fatigue. Other works, such as [6], have
employed even lower frequencies. PW and amplitude were set
according to the control strategies depicted in Figure 3. Reference cycling cadence was also kept constant during all
sessions at 50 r/min. The FES cycling performance was
increased using alternate sessions for muscle strengthening
and skill development. During this phase, we monitored
physiological variables such as heart rate, arterial blood
pressure, lactate level, and blood glucose level. Nevertheless,
because the pilot could not quantify his perception of his
effort, we did not have all the information necessary to estimate the precise exercise level. Based on lactate level measurements, which featured readings above 7 mmol/L during

Figure 6. The FES cycling competition during Cybathlon 2016,
particularly featuring the EMA tricycle and its pilot, Estevão
Lopes.

only four sessions, we concluded that a moderate to intense
level of exercise was maintained.
Competition and Follow-Up
Weeks before the competition, our pilot strengthened his performance, moving from stationary to mobile cycling. Because
we had little knowledge of other teams' performances, there
was no basis for preparing his strategy for the race. In his last
preparation sessions, he attempted to optimize the timing for
FES intensity and tricycle gear modulation, focusing on
increasing his speed at the race's start.
On race day, the system did not present any malfunction
that affected our pilot's performance (Figure 6). Using three
stimulation channels on
each leg (one for the gluteus and two for the
The FES cycling
quadriceps), he was one
of the fastest pilots in the
performance was increased
initial part of the race.
The hamstrings were not
using alternate sessions for
stimulated during the
competition, as they premuscle strengthening and
sented an intense and
increased number of
skill development.
spasms on that day, particularly when stimulated. In the first round, our pilot raced the BerkelBike, UK,
which finished second in the overall competition. Our
pilot completed the first 25 m in 13.1 s, while his opponent
achieved the same distance in 14.6 s. During the race, stimulation parameters were set to 30 Hz and 500 μs. The FES
amplitude was directly controlled by the pilot. From initial
values of 40 and 50 mA to the gluteus and quadriceps
muscles, these amplitudes reached a maximum of 86 and
96 mA, respectively.
During the first lap, video footage indicates that our pilot
suffered a knee fracture. Because this unfortunate event was
not detected at the time, he continued racing, eventually
participating in the final round of the competition. Although
DECEMBER 2017

IEEE ROBOTICS & AUTOMATION MAGAZINE

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - December 2017