首页|基于厚尾双学生氏t分布的非线性状态空间系统鲁棒辨识方法

基于厚尾双学生氏t分布的非线性状态空间系统鲁棒辨识方法

扫码查看
原文链接
  • 万方数据
  • 维普
状态空间模型作为一种常见且重要的模型结构在自动化领域有着广泛的应用,本文针对异常值干扰下的非线性状态空间系统辨识问题开展研究.与现有的辨识方法不同,本文充分考虑了状态转移过程和输出量测过程均受到异常值干扰的情况,提出了一种更加全面的鲁棒辨识算法.首先利用两个相互独立的学生氏t分布分别对状态噪声和输出噪声进行建模以保障算法的鲁棒性;其次利用粒子平滑算法估计状态变量的后验概率分布以解决状态未知问题;最后利用期望最大化算法实现未知参数估计.在算法实现过程中使用了学生氏t分布表达式的数学分解,这样做的好处是:(1)更加有利于算法的推导和实现;(2)更清晰地解释了算法的鲁棒性能.并且本文通过数值算例和应用算例验证了该方法的有效性.
Robust Identification of Nonlinear State-Space System Based on Dual Heavy-Tailed Noise Distributions
The state space model is a common and important model structure for automation and control.In this pa-per,the robust identification of nonlinear state-space model corrupted by outliers is investigated.The outliers imposed on both the state transition process and the output measurement process are considered and a more comprehensive and robust identification algorithm is proposed.To ensure the robustness of the proposed algorithm,two independent heavy-tailed Stu-dent's t-distributions are used to describe the state noise and the output noise,respectively.Then the particle smoothing method is applied to estimate the posterior distribution of the unknown states.Finally,the expectation maximization algo-rithm is used to realize the parameter estimation problem.The mathematical decomposition of the Student's t-distribution is employed in the identification process which brings two main advantages:(1)facilitating the derivation and implementation of the proposed algorithm;(2)providing a more clearer explanation of the robustness of the algorithm.The usefulness of the proposed algorithm is demonstrated via the numerical and mechanical examples.

nonlinear state-space modelrobust system identificationStudent's t-distributionparticle smootherex-pectation maximization algorithm

刘鑫、海洋、代伟

展开 >

中国矿业大学人工智能研究院,江苏 徐州 221116

中国矿业大学信息与控制工程学院,江苏 徐州 221116

非线性状态空间系统 鲁棒辨识 学生氏t分布 粒子平滑 期望最大化算法

国家自然科学基金国家自然科学基金国家重点研发计划中国博士后基金

62103134623733612022YFB33047002023M743776

2024

10.12263/DZXB.20230957

电子学报
中国电子学会

电子学报

CSTPCD北大核心
影响因子：1.237
ISSN：0372-2112
年,卷(期)：2024.52(9)