摘要
利用神经网络对非线性映射的逼近能力，通过采用一步超前预测控制性能指标及网络模型局部线性化的思想，给出了一个显式的控制律和相应的自适应控制算法．仿真结果表明了该控制算法的有效性．
Based on the nonlinear approximation capability of the backpropagation networks,an explicit control law and an adaptive control algorithm for unknown multivariables discrete-time dymamic systems described by an input/output model are presented by using Clarke's one-step-ahead predictive control performance index and the so-called linearization techniques of the local network predictor models. The simulation results showed the effectiveness of the control scheme proposed for tracking control problems of nonlinear or/and time-varying plants.
出处
《东北大学学报（自然科学版）》
EI
CAS
CSCD
1995年第3期257-262,共6页
Journal of Northeastern University(Natural Science)
基金
国家自然科学基金
关键词
非线性系统
自适应控制
神经网络
学习算法
multivariables systems,adaptive control,neural networks,learning algorithm.
