计算机科学
非线性系统
控制理论(社会学)
控制器(灌溉)
集合(抽象数据类型)
人工神经网络
线性模型
基础(线性代数)
数学优化
控制(管理)
人工智能
数学
机器学习
物理
几何学
生物
程序设计语言
量子力学
农学
作者
Weiqiang Tang,Wenkun Long,Tianpeng Xu,Haiyan Gao
标识
DOI:10.3103/s0146411622060098
摘要
Abstract In multiple model control, the model set is very important to the control performance. A novel multiple model control algorithm based on dynamic optimization is proposed for a nonlinear system with jumping parameters. The model set includes linear and nonlinear models. The linear models include some fixed and adaptive models which approximate the linear part of the system. Specially, the fixed models are constantly optimized using the new available data. For the nonlinear part of the system, a neural network is used to describe it. On this basis, the controller set corresponding to multiple models is designed according to the principle of asymptotic tracking. In practice, the switching between controllers is determined by a performance index. The results show that the optimization of the model set is helpful to improve the system control performance. The effectiveness of the proposed method is verified by the simulation results.
科研通智能强力驱动
Strongly Powered by AbleSci AI