Scholarlink

<p>In this paper, we propose an innovative control approach for human–exoskeleton dynamics that integrates a nonlinear extended disturbance observer (EDOB) with a sliding mode control (SMC) featuring a PID sliding surface. The observer plays a dual role by both rejecting lumped disturbances and providing sensorless interaction torque estimates that drive an admittance layer, removing the need for force/torque sensors. This approach aims to optimize the admittance control of a lower-limb exoskeleton during the swing phase, addressing critical issues such as varying interaction forces and dynamic uncertainties. The target application is a pre-gait, non-weight-bearing “air-walking” scenario that isolates swing-phase control without stabilizing ground-reaction forces. The EDOB is designed to estimate and compensate for external disturbances, such as interaction forces, in real time, eliminating the need for additional sensors like load cells. This simplification makes the exoskeleton more lightweight, cost-effective, and easier to manufacture. Meanwhile, the SMC with a PID sliding surface minimizes chattering and improves trajectory tracking precision. The stability and robustness of the proposed controller are mathematically proven, and extensive simulations in MATLAB validate its superior performance compared to traditional PID and SMC controllers. The findings highlight the potential of this strategy to advance the development of adaptive and robust control systems for wearable robotics in rehabilitation applications, offering practical benefits for both users and manufacturers.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Nonlinear disturbance observer-based PID sliding mode admittance control of a lower-limb wearable robot considering a human–exoskeleton coupling dynamics in a swing phase

  • 期刊论文

  • 发表日期：

    2026年2月25日

查看全文
  • Mohamed Elstohy,
  • Kwan-Woong Gwak

摘要

In this paper, we propose an innovative control approach for human–exoskeleton dynamics that integrates a nonlinear extended disturbance observer (EDOB) with a sliding mode control (SMC) featuring a PID sliding surface. The observer plays a dual role by both rejecting lumped disturbances and providing sensorless interaction torque estimates that drive an admittance layer, removing the need for force/torque sensors. This approach aims to optimize the admittance control of a lower-limb exoskeleton during the swing phase, addressing critical issues such as varying interaction forces and dynamic uncertainties. The target application is a pre-gait, non-weight-bearing “air-walking” scenario that isolates swing-phase control without stabilizing ground-reaction forces. The EDOB is designed to estimate and compensate for external disturbances, such as interaction forces, in real time, eliminating the need for additional sensors like load cells. This simplification makes the exoskeleton more lightweight, cost-effective, and easier to manufacture. Meanwhile, the SMC with a PID sliding surface minimizes chattering and improves trajectory tracking precision. The stability and robustness of the proposed controller are mathematically proven, and extensive simulations in MATLAB validate its superior performance compared to traditional PID and SMC controllers. The findings highlight the potential of this strategy to advance the development of adaptive and robust control systems for wearable robotics in rehabilitation applications, offering practical benefits for both users and manufacturers.