摘要
针对单一PID方法控制飞行模拟机电动操纵负荷系统时存在非线性、外力干扰、多余力及外环震荡等问题,根据FCMAC可学习任意多维非线性映射,FCMAC与PID并行控制器鲁棒性强和BP神经网络前馈补偿器具有自适应性、可有效降低多余力影响的特点,采用了由FCMAC与PID并行控制器和BP神经网络前馈补偿器组成的复合神经网络控制方法,解决了单一PID控制方法中存在的问题。对基于复合神经网络控制方法的系统进行了建模与仿真,仿真结果表明该方法使模型力与输出力偏差由1.1N降低到0.1N,干扰信号作用下力偏差保持在1.5%以内,将超调量由35%降低到2%,稳定时间由0.7s缩短到0.05s。
There are problems such as nonlinearity,disturbance,extraneous force and oscillation of exter-nal ring in a flight simulator electric control loading system controlled by single PID. The parallel control-ler neural networks has the feature of robust,fuzzy cerebellar model articulation controller （FCMAC） can study any nonlinear map. Also BP neural networks compensator has the character of being adaptive and the ability to reduce the influence of extraneous force. Therefore a composite neural network control algo-rithm was proposed,which consists of FCMAC neural networks and PID parallel controller,and back propagation （BP） neural networks. The control system was modeled and simulated based on the compos-ite neural network. The results show that the deviation between model torque and output torque is induced from 1. 1 N to 0. 1 N,and the deviation in disturbing signal stays the range of 1. 5% . Also the overshoot is induced from 35% to 2% ,and stabilization time is shortened from 0. 7 s to 0. 05 s.
出处
《电机与控制学报》
EI
CSCD
北大核心
2010年第5期73-78,共6页
Electric Machines and Control
基金
中国民用航空总局科技项目(MHR0703)
