Scholarlink

<p>This paper presents a novel control framework that integrates filtered dynamic inversion (FDI) with model predictive control (MPC) to address the challenges of controlling uncertain linear systems subjected to unknown and unmeasured disturbances while ensuring input constraints are satisfied. Traditional FDI provides robust disturbance rejection with minimal model knowledge but lacks mechanisms to handle actuator constraints, risking control saturation. Meanwhile, MPC effectively enforces input constraints but relies on an accurate system model. To bridge this gap, we propose an FDI–MPC framework that leverages FDI’s disturbance rejection while incorporating MPC to regulate the control input within feasible limits. A key contribution of this work is the development of a novel constraint mapping algorithm, derived from the FDI low-pass filter dynamics, which dynamically transforms actuator constraints into equivalent bounds on the MPC virtual control input. This enables the use of the FDI low-pass filter while enforcing input constraints and preserving disturbance rejection. The proposed method’s effectiveness is validated through simulations, showing its ability to follow commands, reject disturbances, and prevent saturation, making it suitable for real-world applications with uncertain dynamics.</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

Filtered dynamic inversion for input constrained uncertain linear systems with unknown and unmeasured disturbances

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月21日

查看全文
  • Arash Rahmanidehkordi,
  • Amirhossein Ghasemi

摘要

This paper presents a novel control framework that integrates filtered dynamic inversion (FDI) with model predictive control (MPC) to address the challenges of controlling uncertain linear systems subjected to unknown and unmeasured disturbances while ensuring input constraints are satisfied. Traditional FDI provides robust disturbance rejection with minimal model knowledge but lacks mechanisms to handle actuator constraints, risking control saturation. Meanwhile, MPC effectively enforces input constraints but relies on an accurate system model. To bridge this gap, we propose an FDI–MPC framework that leverages FDI’s disturbance rejection while incorporating MPC to regulate the control input within feasible limits. A key contribution of this work is the development of a novel constraint mapping algorithm, derived from the FDI low-pass filter dynamics, which dynamically transforms actuator constraints into equivalent bounds on the MPC virtual control input. This enables the use of the FDI low-pass filter while enforcing input constraints and preserving disturbance rejection. The proposed method’s effectiveness is validated through simulations, showing its ability to follow commands, reject disturbances, and prevent saturation, making it suitable for real-world applications with uncertain dynamics.