Scholarlink

<p>Precise trajectory tracking is critical for quadrotors operating in complex environments but is often degraded by coupled modeling uncertainties, parameter variations, and external disturbances. To address the limited disturbance adaptivity and residual chattering in existing robust controllers, this paper proposes an adaptive sliding-mode control scheme augmented with an extreme learning machine (ASMC-ELM). The quadrotor dynamics are derived via the Lagrangian method and decomposed into position and attitude subsystems using time-scale separation. A PID-type sliding surface with a fast terminal reaching law is designed to accelerate convergence and suppress chattering, while an ELM is embedded to online estimate and compensate the lumped disturbance in real time. Simulations and flight tests verify the effectiveness of the proposed method, achieving faster transient response, smaller steady-state tracking errors, and reduced control-input chattering compared with baseline controllers.</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

Adaptive Sliding-Mode Trajectory Tracking Control for Quadrotors Using ELM-Based Disturbance Estimation

  • 期刊论文

  • 发表日期：

    2026年4月21日

查看全文
  • Li Ding,
  • Fengze Jiang,
  • Hongtao Wu,
  • Yangmin Li

摘要

Precise trajectory tracking is critical for quadrotors operating in complex environments but is often degraded by coupled modeling uncertainties, parameter variations, and external disturbances. To address the limited disturbance adaptivity and residual chattering in existing robust controllers, this paper proposes an adaptive sliding-mode control scheme augmented with an extreme learning machine (ASMC-ELM). The quadrotor dynamics are derived via the Lagrangian method and decomposed into position and attitude subsystems using time-scale separation. A PID-type sliding surface with a fast terminal reaching law is designed to accelerate convergence and suppress chattering, while an ELM is embedded to online estimate and compensate the lumped disturbance in real time. Simulations and flight tests verify the effectiveness of the proposed method, achieving faster transient response, smaller steady-state tracking errors, and reduced control-input chattering compared with baseline controllers.