Abstract:
Wearable devices such as ankle exoskeletons have demonstrated the ability to enhance human
mobility and reduce the biological efforts of human locomotion. Previous studies have
shown that incorporating clutches and elastic elements in parallel with the ankle muscles
can offload substantial muscle force and therefore reduce the metabolic cost of walking.
However, the optimal assistance timing with clutched elastic exoskeletons remains unclear.
The purpose of this study was to evaluate the effect of different assistance onset timings on
human gait mechanics and energetics by regulating the activation of a controllable clutch
in an elastic ankle exoskeleton. This mechanism uses the principle of controlled energy
storage and release and has been designed to optimise the functionality of a passive elastic
exoskeleton. For this purpose, I designed and developed a semi-passive clutched spring ankle
exoskeleton that enables us to study different assistance onset timings on healthy individuals’
gait biomechanics and energetics across different walking speeds.
The pneumatic clutch-spring design embodied the high force-to-mass and force-to-power
ratio with much less mechanical complexity than other existing clutch mechanisms while also
achieving rapid response times and high fidelity. To demonstrate the functionality, validity,
and reliability of the system, the device was tested on a healthy participant (female; 72 kg,
172 cm; 29-year old) while walking at 1.25 m.s−1 with the clutch engaged at three activation
angle points (P0, P1, and P2) during the early stance. The results from walking tests revealed
the ability of the system to vary the actuation timing reliably and to change the assistance
profile across the gait cycle with high accuracy.
Additionally, tests on a larger sample of participants have been performed to study the
performance of the clutched elastic ankle exoskeleton across a different range of human walking speeds and to understand how the actuation timing and walking speed influence the
participant’s biomechanics and energetics. Ten able-bodied adults (n = 10; 78.5±5.5 kg
body mass, 177 ± 0.5 cm; 29 ± 5 years old; means ± SD) participated in the study and
completed a series of walking trials at three walking speeds (1.25, 1.50, and 1.75 m.s−1
) on
a force-instrumented treadmill. The participants read and signed a written consent prior to
participation. The participants walked with the device unassisted (kexo = 0 N.m.rad1
) and
then with the exoskeleton stiffness of kexo = 105 N.m.rad1
(assisted) with the clutch engaged
at three points during the early stance (P0: heel-strike, P1: foot-flat, and P2: mid-stance) and
disengaged at the same angle during the late stance. In all trials, lower limb joint kinematics,
muscle activation, and whole-body metabolic rate were measured.
This study found that using a controllable clutched elastic ankle exoskeleton can improve
walking economy up to 8% as compared to unassisted walking with the exoskeleton across
all speeds. The results also suggest that controlling the timing of assistance in clutched
elastic ankle exoskeletons can influence the assistive performance and further improve gait
efficiency. This emphasises the importance of using a controllable clutch along with passive
elastic ankle exoskeletons to optimise assistance during human gait.
This thesis gives insight into how adjusting exoskeleton characteristics (i.e., assistance
timing and magnitude) can improve walking economy and demonstrates that the degree to
which individual users will benefit from exoskeleton assistance is highly dependent on their
specific needs, expected walking conditions, and the choice of outcome measures. I hope the
outcome of this work contributes to a better understanding of human physiological responses
and a shift towards individualised wearable robotic design and control.