三通路反射控制的人体姿势平衡预测仿真

Three-Pathway Reflex Control for Predictive Simulation of Human Postural Balance

摘要: 反馈控制模型已被用于人类姿势控制的运动模拟研究。本研究基于肌肉骨骼建模与控制，改进了一种三通路反馈控制器，用于预测不同感觉运动时延与扰动幅度下的人体姿势平衡响应。基于改编的gait10dof18musc模型，在100、150和200 ms三种反馈延迟和0 mm~80 mm的不同幅度下对扰动平衡进行了模拟。结果显示，较长时延导致更高的峰值肌肉激活；随着扰动幅度增大，踝与膝关节的活动范围增加而骨盆与髋关节活动范围呈下降趋势；多块关键肌在不同时延间的峰值激活存在显著差异；此外，相较原控制器，改进的反馈控制器在更广的时延范围和扰动幅度内维持了更小的CoM与CoP偏差，四对主要的拮抗肌共收缩指数明显降低。这些发现表明，该控制器在更广时延和扰动范围内鲁棒性更好，在扰动跟踪与姿态稳定方面更为有效。

Abstract: Feedback-control models have been widely used to simulate human postural control. Building on musculoskeletal modeling and control, this study improves a three-path feedback controller to predict human postural responses under varying sensorimotor delays and perturbation amplitudes. Using a modified two-dimensional gait10dof18musc model, we simulated perturbed balance at feedback delays of 100, 150, and 200 ms and amplitudes of 0 mm~80 mm. The results show that longer delays lead to higher peak muscle activations; as perturbation amplitude increases, the ranges of motion at the ankle and knee increase, whereas those of the pelvis and hip decrease; and several key muscles exhibit significant between-delay differences in peak activation. Moreover, relative to the original controller, the improved controller maintains smaller CoM-CoP deviations over a broader range of delays and amplitudes and markedly reduces the co-contraction index for four principal antagonist pairs. These findings indicate that the improved controller achieves greater robustness across delays and perturbations and is more effective in perturbation tracking and postural stabilization.