摘要
针对一类未知非线性系统的建模与控制问题,采用基于神经网络的多步预测控制算法进行了仿真研究,仿真对象选取控制工程界常用的单摆试验装置.该算法的实现步骤为:获取系统开环试验数据;辨识神经网络正向动态模型;设计非线性优化控制律.仿真研究结果表明,所提出的建模与控制方法是有效的.
To achieve the effective modeling and control strategy of a class of unknown nonlinear system, the in-depth theoretical exploration is carried through with a single pendulum as a study case. The application of NN-based method to identify the unknown nonlinear system, and the NN-based nonlinear multi-steps predictive control method are presented, which can be implemented in three steps: obtaining the open-loop test data, NN identification of the forward dynamic model, and implementing the nonlinear optimal control law. The simulation result shows that the modeling and control method is effective.
出处
《海军工程大学学报》
CAS
2004年第4期53-56,共4页
Journal of Naval University of Engineering
基金
海军工程大学科学研究基金资助项目(HGDJJ01005)
关键词
未知非线性系统
神经网络
建模与控制
unknown nonlinear system
neural networks
modeling and control
