Gait disabilities empowered intensive research on the field of human-robot interaction to promote effective gait rehabilitation. Assist-as-needed strategies are becoming prominent, appealing to the users’ participation in their rehabilitation therapy. This study proposes and assesses the biomechanical effects of an adaptive impedance control strategy that innovatively allows adaptability in interaction-based stiffness and gait trajectory towards a fully assist-as-needed therapy. By modulating the interaction-based stiffness per gait phase, we hypothesize that the strategy appeals to a symbiotic human-orthotic cooperation, augmenting the user’s muscular activity. The interaction stiffness was estimated by modelling the human-orthosis interaction torque vs angle curve with a linear regression model. The strategy also allows for real-time trajectory adaptations at different gait phases to fulfil the users’ needs. The biomechanical assessment of the impedance-controlled ankle orthosis involved eight healthy volunteers walking at 1.0 and 1.6 km/h. The results revealed a stronger muscular activation regarding the non-assisted leg for the gastrocnemius lateralis (increment ratio ≥ 1.0 for both gait speeds) and for the tibialis anterior muscle (increment ratio ≥ 1.0 for 1.6 km/h). The strategy guided users successfully on a healthy gait pattern while allowing deviations (median error < 5.0°) given the users’ intention weighted by interaction stiffness. Findings showed the relevance for adapting gait trajectory as users prefer higher trajectories as the speed increases. No significant temporal variations or neither knee angular compensations were observed (p value ≥0.11). Overall results support that this strategy may be applied for intensity-adapted gait training, allowing different human-robot compliant levels.

步态障碍促进了人机交互领域的深入研究，以促进有效的步态康复。按需辅助策略正变得越来越突出，吸引用户参与他们的康复治疗。这项研究提出并评估了自适应阻抗控制策略的生物力学效应，该策略创新地允许基于交互的刚度和步态轨迹的适应性，以实现按需治疗。通过调节每个步态阶段基于交互的刚度，我们假设该策略吸引了人类与矫形器的共生合作，增强了用户的肌肉活动。通过使用线性回归模型对人体-矫形器相互作用扭矩与角度曲线进行建模来估计相互作用刚度。该策略还允许在不同步态阶段进行实时轨迹调整，以满足用户的需求。阻抗控制踝关节矫形器的生物力学评估涉及八名健康志愿者以 1.0 和 1.6 公里/小时的速度行走。结果显示，非辅助腿的肌肉激活更强。腓肠肌（两种步态速度的增量比 ≥ 1.0）和胫骨前肌（1.6 km/h 的增量比 ≥ 1.0）。该策略成功地引导用户进入健康的步态模式，同时允许偏差（中值误差 < 5.0°），给定用户意图由交互刚度加权。调查结果显示了适应步态轨迹的相关性，因为随着速度的增加，用户更喜欢更高的轨迹。没有观察到显着的时间变化或膝关节角度补偿（p值≥0.11）。总体结果支持该策略可应用于适应强度的步态训练，允许不同的人机合规水平。