摘要
为了实现对非线性广义预测控制算法的仿真研究,采用神经网络建立非线性预测模型,在其工作点附近线性化,并对线性化模型进行广义预测控制,利用单片输出BP网络所辨识的非线性模型进行分析,提出了一种基于非线性广义预测前馈补偿控制律的补偿算法,改善了线性化所带来的模型失配误差。最后通过Matlab进行了仿真。仿真结果验证了算法的有效性。
Aimed at realizing the simulating research of the nonlinear GPC algorithm, neural network control（NNC） was adopted to build nonlinear predictive model that was linearized around working point. Then the GPC was introduced to control liaearized model and the nonlinear model that BP neural network of single output reeogniped was used to analyze. A compensation algorithm based on nonlinear GPC and frnnt-feedbaek compensation control theory was presented to improve the model inaccuracy. The Matlab simulation results indicate the availability of the developed algorithm.
出处
《机电工程》
CAS
2008年第11期101-104,共4页
Journal of Mechanical & Electrical Engineering
关键词
广义预测控制
神经网络控制
模型失配
general predictive control（ GPC）
neural network control
model inaccuracy
