摘要
基于神经网络的预测控制器由神经单元自适应PID控制器和基于神经网络的Smith预估器组成。预估器对输出进行多步预测,使控制器超前动作以消除时滞对系统的影响。自适应PID控制器通过有监督的Hebb学习算法实现其权值的调节,同时通过权系数的在线调整实现自适应控制,提高自适应能力。
Predictive controller based on neural network is composed of neural adaptive PID controller and Smith predictive controller. The predictor controller makes multi-step prediction for the output so that the predictor operates in advance to eliminate the time delay effect to the system. The weights was adjusted by adaptive PID controller through the supervised Hebb learning algorithm, and adaptive control was realized through regulating the coefficient of weight on-line, to improve the adaptive ability.
出处
《兵工自动化》
2005年第4期51-52,共2页
Ordnance Industry Automation