Scholarlink

<p>This study presents the design and comparative analysis of advanced nonlinear sliding mode controllers for formation control of multiple unmanned helicopters under external disturbances. Four different control strategiesFirst-Order SMC, Terminal SMC, Non-Singular Super-Twisting SMC (NST-SMC), and Reinforcement Learning–Enhanced NST-SMC (RL–NST-SMC) are developed and evaluated. The leader–follower formation structure is adopted, and the proposed controllers are tested in multiple missions involving spiral and S-shaped trajectories with varying numbers of helicopters. Simulation results demonstrate that the RL–NST-SMC achieves the best overall performance, offering fast convergence, minimal tracking error, smooth control effort, and high robustness against disturbances. Compared with conventional SMC methods, the RL–NST-SMC reduces chattering, overshoot, and control energy consumption while maintaining precise trajectory tracking. Sensitivity analyses further confirm the superior adaptability and parameter robustness of the proposed intelligent controller. These results validate the RL–NST-SMC as a promising approach for high-precision cooperative flight control of unmanned helicopter formations.</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

Reinforcement learning–enhanced non-singular super-twisting sliding mode control for robust formation flight of unmanned helicopters

  • 期刊论文

  • 发表日期：

    2026年6月29日

查看全文
  • A. Toloei,
  • F. Ghaderi

摘要

This study presents the design and comparative analysis of advanced nonlinear sliding mode controllers for formation control of multiple unmanned helicopters under external disturbances. Four different control strategiesFirst-Order SMC, Terminal SMC, Non-Singular Super-Twisting SMC (NST-SMC), and Reinforcement Learning–Enhanced NST-SMC (RL–NST-SMC) are developed and evaluated. The leader–follower formation structure is adopted, and the proposed controllers are tested in multiple missions involving spiral and S-shaped trajectories with varying numbers of helicopters. Simulation results demonstrate that the RL–NST-SMC achieves the best overall performance, offering fast convergence, minimal tracking error, smooth control effort, and high robustness against disturbances. Compared with conventional SMC methods, the RL–NST-SMC reduces chattering, overshoot, and control energy consumption while maintaining precise trajectory tracking. Sensitivity analyses further confirm the superior adaptability and parameter robustness of the proposed intelligent controller. These results validate the RL–NST-SMC as a promising approach for high-precision cooperative flight control of unmanned helicopter formations.