摘要
本文基于ＢＰ结构神经网络，对系统的建模误差进行预测，并将其与模型预测相结合构成广义预测控制算法，目的在于抑制模型失配的影响，增强广义预测控制的鲁棒性，仿真结果表明了这一算法的有效性．
The error prediction in building the system model based on the BP network is proposed inthis paper, and this prediction is combined with the model prediction to form the generalized predictivecontrol (GPC). This is aimed at offsetting the effect of the model mismatch, and strengthening the robustness of the GPC algorithm. The simulation results show this control method is effective.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
1996年第5期677-680,共4页
Control Theory & Applications
关键词
神经网络
预测控制
误差修正
鲁棒性
neural network
predictive control
error correction
model mismatch
robustness
