Scholarlink

<p>This study proposes an improved adaptive fault-tolerant control (FTC) scheme for switched nonlinear systems with multichannel actuators and deferred constraints. First, a novel multivariable lemma for type-<i>B</i> Nussbaum functions is developed, in which the monotonicity requirement on the Nussbaum argument is removed under time-varying gains. Second, inspired by the concept of deferred constraints, an enhanced tracking control method is proposed, which achieves excellent tracking performance without imposing strict initial conditions. Specifically, the tracking error rapidly converges to a specified region within a preset time without noticeable overshoot and asymptotically converges to zero. Third, a new fault-tolerant controller is designed, which consists of a control protocol and a learning-based switching mechanism. Finally, the proposed mechanism enables fault detection and automatic switching of working modes without human intervention in a reasonably efficient manner. The proposed FTC architecture does not need an additional fault detection and/or isolation unit, reducing the risk of false and missed alarms. Moreover, the simplified design framework avoids filtering techniques and prevents an “explosion of complexity”. All closed-loop signals are proven to be semiglobally uniformly ultimately bounded (SUUB) under arbitrary switching, and excellent tracking performance is realized. Simulation results confirm the superiority of the proposed control method.</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

Fault-tolerant control for switched nonlinear systems with multichannel actuators and deferred constraints via a learning-based switching mechanism

  • 期刊论文

  • 发表日期：

    2026年5月26日

查看全文
  • Wei Qian,
  • Quan Wan,
  • Bo Shen

摘要

This study proposes an improved adaptive fault-tolerant control (FTC) scheme for switched nonlinear systems with multichannel actuators and deferred constraints. First, a novel multivariable lemma for type-B Nussbaum functions is developed, in which the monotonicity requirement on the Nussbaum argument is removed under time-varying gains. Second, inspired by the concept of deferred constraints, an enhanced tracking control method is proposed, which achieves excellent tracking performance without imposing strict initial conditions. Specifically, the tracking error rapidly converges to a specified region within a preset time without noticeable overshoot and asymptotically converges to zero. Third, a new fault-tolerant controller is designed, which consists of a control protocol and a learning-based switching mechanism. Finally, the proposed mechanism enables fault detection and automatic switching of working modes without human intervention in a reasonably efficient manner. The proposed FTC architecture does not need an additional fault detection and/or isolation unit, reducing the risk of false and missed alarms. Moreover, the simplified design framework avoids filtering techniques and prevents an “explosion of complexity”. All closed-loop signals are proven to be semiglobally uniformly ultimately bounded (SUUB) under arbitrary switching, and excellent tracking performance is realized. Simulation results confirm the superiority of the proposed control method.