Scholarlink

<p>To address the trajectory tracking accuracy degradation of the picking robotic arm in mountainous orchards due to strong nonlinear dynamics, parameter uncertainties, and multiple time-varying disturbances in complex unstructured environments, this paper proposes a nonsingular fast terminal sliding mode control method with radial basis function neural network (RBFNN) compensation and a dynamic adaptation mechanism. First, the three-channel independent RBFNN is designed to online approximate the total uncertainty disturbances, and an adaptive mechanism is introduced to adaptively tune the weights online. Next, the innovative dynamic adaptive adjustment mechanism for the nonsingular fast terminal sliding surface gain parameters is constructed. A real-time error-based adaptive law is designed to adjust the key gains: enhancing convergence speed during the large-error phase and actively suppressing chattering during the small-error phase, effectively mitigating the inherent contradiction between convergence speed and steady-state accuracy under time-varying disturbances in traditional sliding mode control. The finite-time stability of the closed-loop system under uncertainty disturbances is rigorously proven using Lyapunov theory. Finally, the proposed method is evaluated on three-link and five-link manipulators. Comparative simulations on the 3-DOF system under composite disturbances demonstrate faster convergence and active impact compensation. Furthermore, 5-DOF system simulations verify high-precision tracking under disturbances, confirming its overall effectiveness for orchard harvesting.</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

Radial Basis Function Neural Network-Based Adaptive Nonsingular Fast Terminal Sliding Mode Control for Trajectory Tracking of Robotic Manipulators

  • 期刊论文

  • 发表日期：

    2026年4月29日

查看全文
  • Shengyang Tao,
  • Yuanping Su,
  • Dongyun Yu

摘要

To address the trajectory tracking accuracy degradation of the picking robotic arm in mountainous orchards due to strong nonlinear dynamics, parameter uncertainties, and multiple time-varying disturbances in complex unstructured environments, this paper proposes a nonsingular fast terminal sliding mode control method with radial basis function neural network (RBFNN) compensation and a dynamic adaptation mechanism. First, the three-channel independent RBFNN is designed to online approximate the total uncertainty disturbances, and an adaptive mechanism is introduced to adaptively tune the weights online. Next, the innovative dynamic adaptive adjustment mechanism for the nonsingular fast terminal sliding surface gain parameters is constructed. A real-time error-based adaptive law is designed to adjust the key gains: enhancing convergence speed during the large-error phase and actively suppressing chattering during the small-error phase, effectively mitigating the inherent contradiction between convergence speed and steady-state accuracy under time-varying disturbances in traditional sliding mode control. The finite-time stability of the closed-loop system under uncertainty disturbances is rigorously proven using Lyapunov theory. Finally, the proposed method is evaluated on three-link and five-link manipulators. Comparative simulations on the 3-DOF system under composite disturbances demonstrate faster convergence and active impact compensation. Furthermore, 5-DOF system simulations verify high-precision tracking under disturbances, confirming its overall effectiveness for orchard harvesting.