Gait Asymmetry from Unilateral Weakness and Improvement With Ankle Assistance: a Reinforcement Learning based Simulation Study

TL;DR

A reinforcement learning (RL) based musculoskeletal simulation framework supports controlled evaluation of impairment severity and assistive strategies, and provides a basis for future validation in human experiments.

Abstract

Unilateral muscle weakness often leads to asymmetric gait, disrupting interlimb coordination and stance timing. This study presents a reinforcement learning (RL) based musculoskeletal simulation framework to (1) quantify how progressive unilateral muscle weakness affects gait symmetry and (2) evaluate whether ankle exoskeleton assistance can improve gait symmetry under impaired conditions. The overarching goal is to establish a simulation- and learning-based workflow that supports early controller development prior to patient experiments. Asymmetric gait was induced by reducing right-leg muscle strength to 75%, 50%, and 25% of baseline. Gait asymmetry was quantified using toe-off timing, peak contact forces, and joint-level symmetry metrics. Increasing weakness produced progressively larger temporal and kinematic asymmetry, most pronounced at the ankle. Ankle range of motion symmetry degraded from near-symmetric behavior at 100% strength (symmetry index, SI = +6.4%; correlation r=0.974) to severe asymmetry at 25% strength (SI = -47.1%, r=0.889), accompanied by a load shift toward the unimpaired limb. At 50% strength, ankle exoskeleton assistance improved kinematic symmetry relative to the unassisted impaired condition, reducing the magnitude of ankle SI from 25.8% to 18.5% and increasing ankle correlation from r=0.948 to 0.966, although peak loading remained biased toward the unimpaired side. Overall, this framework supports controlled evaluation of impairment severity and assistive strategies, and provides a basis for future validation in human experiments.

Figures (5)

• Figure 1: Reinforcement learning framework for human--exoskeleton multi-agent training, adapted from tan2025myoassist. Stage 1 trains a baseline human locomotion policy for normal walking. Stage 2 performs transfer learning under progressive unilateral weakness; joint human--exoskeleton training for right-ankle assistance was performed only at $\alpha=0.5$ using PPO schulman2017proximal.
• Figure 2: Joint angle trajectories during steady-state walking under progressive unilateral right-leg muscle weakness. Panels correspond to different muscle strength levels: (a) 100% strength ($\alpha = 1.00$), (b) 75% strength ($\alpha = 0.75$), (c) 50% strength ($\alpha = 0.50$), and (d) 25% strength ($\alpha = 0.25$). Solid and dashed lines denote right and left legs, respectively; shaded regions indicate $\pm$1 SD across gait cycles. Toe-off is marked by vertical lines (solid right, dashed left).
• Figure 3: Ground contact force profiles during stance under progressive unilateral right-leg muscle weakness (see panel labels for $\alpha$). Blue/red: left/right foot; shaded: $\pm$1 SD across gait cycles.
• Figure 4: Joint angle trajectories at 50% strength ($\alpha=0.5$) with and without ankle exoskeleton assistance. "50%" denotes unassisted and "50%+Exo" denotes assisted. Solid/dashed lines indicate right/left legs; shaded regions show $\pm1$ SD; vertical lines mark toe-off.
• Figure 5: Learned right-leg ankle exoskeleton torque profile at 50% muscle strength ($\alpha=0.5$). The curve shows the mean commanded ankle torque over one gait cycle for the assisted condition (50%+Exo), with shaded regions indicating $\pm 1$ SD across gait cycles. Vertical line marks right-leg toe-off timing. Positive torque denotes dorsiflexion and negative torque denotes plantarflexion.