Scholarlink

<p>The load-dependent exponential ARX (Exp-ARX) model is a kind of nonlinear combination model. This paper focuses on the parameter estimation for the Exp-ARX model. To overcome the estimation difficulty due to the highly nonlinear relations between the parameters and the model output, the separated idea is used to transform the original optimization problem into a quadratic and nonlinear optimization problem. Applying the hierarchical identification principle, two interactive algorithms are proposed for the Exp-ARX model. In addition, a method based on data optimization of the random gradient forest (RGF) combined with the ability to estimate multiple parameters simultaneously is proposed to solve the problem of locally optimal solutions in traditional optimization methods, which significantly improves the estimation efficiency. The simulation results verify the effectiveness of the proposed algorithms in terms of parameter estimation accuracy and prediction performance. Finally, experiments on piezoelectric ceramic-driven flow control systems demonstrate the excellent performance of the proposed algorithm in dealing with nonlinear characteristics and dynamic fluctuations.</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

Hybrid predictive gradient identification of nonlinear Exp-ARX systems based on the machine learning data preprocessing

  • 期刊论文

  • 发表日期：

    2026年7月27日

查看全文
  • Ya Gu,
  • Dekai Wang,
  • Yonghong Tan,
  • Shiyi Yin

摘要

The load-dependent exponential ARX (Exp-ARX) model is a kind of nonlinear combination model. This paper focuses on the parameter estimation for the Exp-ARX model. To overcome the estimation difficulty due to the highly nonlinear relations between the parameters and the model output, the separated idea is used to transform the original optimization problem into a quadratic and nonlinear optimization problem. Applying the hierarchical identification principle, two interactive algorithms are proposed for the Exp-ARX model. In addition, a method based on data optimization of the random gradient forest (RGF) combined with the ability to estimate multiple parameters simultaneously is proposed to solve the problem of locally optimal solutions in traditional optimization methods, which significantly improves the estimation efficiency. The simulation results verify the effectiveness of the proposed algorithms in terms of parameter estimation accuracy and prediction performance. Finally, experiments on piezoelectric ceramic-driven flow control systems demonstrate the excellent performance of the proposed algorithm in dealing with nonlinear characteristics and dynamic fluctuations.