Scholarlink

<p>This paper presents novel iterative parameter identification methods for multivariable autoregressive output-error moving-average systems. A multivariable system is decomposed into several subsystems to reduce computational complexity, and the corresponding identification models are derived. By means of the maximum likelihood principle, an auxiliary model maximum likelihood gradient-based iterative (AM-ML-GI) algorithm is developed to achieve accurate parameter estimation. To further improve the identification efficiency, a hierarchical auxiliary model maximum likelihood gradient-based iterative (H-AM-ML-GI) algorithm is proposed by using the hierarchical identification principle. A detailed floating point operation analysis shows that the H-AM-ML-GI algorithm achieves higher computational efficiency than the AM-ML-GI algorithm. Simulation results demonstrate the effectiveness of the proposed algorithms in identifying the multivariable system with high parameter estimation accuracy.</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

Hierarchical maximum likelihood gradient-based iterative identification methods for multivariable systems

  • 期刊论文

  • 发表日期：

    2026年1月19日

查看全文
  • Qian Zhang,
  • Ximei Liu

摘要

This paper presents novel iterative parameter identification methods for multivariable autoregressive output-error moving-average systems. A multivariable system is decomposed into several subsystems to reduce computational complexity, and the corresponding identification models are derived. By means of the maximum likelihood principle, an auxiliary model maximum likelihood gradient-based iterative (AM-ML-GI) algorithm is developed to achieve accurate parameter estimation. To further improve the identification efficiency, a hierarchical auxiliary model maximum likelihood gradient-based iterative (H-AM-ML-GI) algorithm is proposed by using the hierarchical identification principle. A detailed floating point operation analysis shows that the H-AM-ML-GI algorithm achieves higher computational efficiency than the AM-ML-GI algorithm. Simulation results demonstrate the effectiveness of the proposed algorithms in identifying the multivariable system with high parameter estimation accuracy.