全文下载:
20250120.pdf
文章编号: 1672-6987(2025)01-0001-16 DOI: 10.16351/j.1672-6987.2025.01.001
丁锋1,2, 徐玲3, 张霄1(1.江南大学 物联网工程学院,江苏 无锡 214122;2.青岛科技大学 自动化与电子工程学院,山东 青岛 266061;3.常州大学 微电子与控制工程学院,江苏 常州 213159)
摘要: 大规模系统结构复杂、时空规模大,变量维数高、信息数据量大、非线性、动态演化性以及各种随机干扰等因素引发的辨识难题,需要基于算法创新和模型创新的研究理念,研究和提出大规模系统的高效辨识方法。主要研究工作包括:1)深入剖析大规模辨识问题复杂性特征,挖掘大规模系统内因和外因引起的关键辨识难点,分析信息数据特征以及随机干扰类型,研究构建模型的优化算法求解的科学计算问题;2)针对白噪声干扰、参数变量维数高,以及大规模计算成本高而效率低的问题,利用递阶计算,研究并行计算平台架构下的大规模系统快速、低成本的高效递阶辨识方法;3)针对有色噪声干扰下大规模系统的非线性、时变性等综合因素影响,利用辅助模型辨识思想、滤波辨识理念,以及自适应补偿手段,研究具有预测和补偿干扰能力的大规模系统高效递阶辨识方法;4)对大规模系统复杂性特征及干扰等因素进行科学抽象,研究大规模系统辨识方法收敛条件,建立收敛理论。
关键词: 参数估计; 系统辨识; 辅助模型辨识; 多新息辨识; 递阶辨识; 耦合辨识; 滤波辨 识; 大规模系统
中图分类号: TP 273 文献标志码: A
引用格式: 丁锋, 徐玲, 张霄.大规模系统高效辨识方法研究[J].青岛科技大学学报(自然科学版), 2025, 46(1): -.
DING Feng, XU Ling, ZHANG Xiao.Research on highly-efficient identification methods for large-scale systems[J].Journal of Qingdao University of Science and Technology(Natural Science Edition), 2025, 46(1): -.
Research on Highly-Efficient Identification Methods for Large-Scale Systems
DING Feng1,2, XU Ling3, ZHANG Xiao1(1.School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, China;2.College of Automation and Electronic Engineering, Qingdao University of Science and Technology, Qingdao 266061, China;3.School of Microelectronics and Control Engineering, Changzhou University, Changzhou 213159, China)
Abstract: For the identification problems of large-scale systems with complex structures, large spatiotemporal scales, high variable dimensions, numerous information and data, nonlinearities, dynamic evolutions and various random disturbances, it is necessary to investigate and propose highly-efficient identification methods for large-scale systems based on the algorithm innovation and model innovation. The main research work includes: 1) deeply analyzing the complexity characteristics of large-scale identification problems, mining the key identification difficulties caused by internal and external factors of large-scale systems, analyzing the characteristics of information data and random interferences, and studying the scientific computing problems solved by the optimization algorithms; 2) using the hierarchical computation to study the rapid, low-cost and highly-efficient hierarchical identification methods of large-scale systems with higher dimensional parameter variables and with white noise interferences under the architecture of parallel computation platforms; 3) proposing highly-efficient hierarchical identification methods of large-scale systems with colored noises through predicting and compensating the interference and using the auxiliary model identification idea, the filtering identification idea and the adaptive compensation for nonlinear and time-varying factors; 4) making the scientific abstract of complex characteristics and interference factors and studying the convergence conditions of large-scale system identification methods and establishing the convergence theory.
Key words: parameter estimation; system identification; auxiliary model identifi cation; multi-innovation identification; hierarchical identification; coupling identification; filtering identification; large-scale system
收稿日期: 2025-01-09
基金项目: 国家自然科学基金项目 (62273167, 61873111).
作者简介: 丁锋,男,博士, “泰山学者” 特聘教授,博士生导师.