摘要
首先提出基于自适应神经元的振动智能ＰＩＤ控制策略及相应高效算法，然后通过数字仿真与模型实验验证了这种算法的有效性。这种方法具有控制器参数少、结构简单、算法收敛速度快、便于实时控制等优点。与传统ＰＩＤ控制相比，控制器参数整定可通过神经网络的自组织来实现。数字仿真与实验结果表明这种方法能够有效地控制动态特性未知。
On the basis of adaptive neurone model and its reinforcement learning rule, an intelligent PID control method is developed with the features such as need of fewer controller parameters, simpler structure, better convergence and easier implementation of realtime control. Compared with traditional PID control strategy, the controller parameters can be adjusted by selforganization of neurone, computer simulations and experiments for controlling the vibration of a flexible structure with five degrees of freedom are implemented. The results show that the method proposed can be effective on the vibration control of any blackbox vibration system with unmeasurable disturbance.
出处
《应用力学学报》
EI
CAS
CSCD
北大核心
1997年第4期64-70,共7页
Chinese Journal of Applied Mechanics
基金
国家自然科学基金
航空科学基金
博士后基金
关键词
结构振动
PID控制
自适应神经元
自组织
: structural vibration, active control, adaptive neurone, PID control.
