Effects of an actuated ankle exoskeleton on walking stability in healthy adults: a controlled laboratory study

$\textbf{Background}$
Ankle exoskeletons are widely used to reduce the metabolic cost of walking, yet their effects on
walking stability during unperturbed gait remain insufficiently understood. Walking stability can be characterized
using complementary measures that capture stride-to-stride variability, global temporal organization, and local
dynamic stability. Understanding how walking with an actuated ankle exoskeleton system influences these different
aspects of gait stability is essential for the safe design and control of wearable robotic devices.

$\textbf{Methods}$
Eighteen healthy adults walked on a treadmill at a constant speed (1.1 m/s) with and without an actuated
bilateral ankle exoskeleton in a randomized crossover design. Spatiotemporal variability was quantified using
coefficients of variation (CoV) of stride length, step width, and stance ratio. Global gait stability was assessed using
detrended fluctuation analysis of stride time. Local dynamic stability was evaluated using maximum Lyapunov
exponent calculated for the trunk, hip, upper leg, lower leg, and foot. Paired-samples two-sided t-tests were used to
... mehr
compare conditions.

$\textbf{Results}$
Walking with the ankle exoskeleton resulted in increased stride-to-stride spatiotemporal variability, reflected
by higher CoV values for stride length (p<0.001) and stance ratio (p=0.005), while mean stride length and step width
remained unchanged. Mean stance ratio was reduced in the exoskeleton condition (p<0.001). Global gait stability did
not differ between conditions, indicating preserved long-range temporal gait organization. Local dynamic stability
increased at the lower leg (p<0.001) and foot (p=0.019) when walking with the exoskeleton.

$\textbf{Conclusions}$
Walking with the actuated ankle exoskeleton alters gait control across multiple levels during steady
walking. While stride-to-stride variability in stride length and stance ratio increased, global gait stability remained
unchanged. Local dynamic stability was increased at the lower leg and foot, suggesting segment-specific effects
of ankle-level assistance close to the assisted joint. However, these findings should be interpreted as the combined
effect of wearing the exoskeleton and receiving active assistance, rather than the isolated effect of plantarflexion assistance. These results provide insight for the design and control of ankle exoskeletons with respect to stabilityrelated effects during walking.