【作者】 吕进； 郭晨； 刘雨；

【Author】 Lü Jin,GUO Chen,LIU Yu(College of Automation and Electric Engineering,Dalian Maritime University,Dalian 116026,China)

【机构】 大连海事大学自动化与电气工程学院；

【摘要】 针对大滞后系统的控制特性,设计一种神经网络二阶导数多步预测模型及其辨识和预测算法,并在此基础上进一步提出基于不完全递归神经网络的预测模型以及模糊小脑模型关节神经网络控制器的控制方案,解决传统自适应控制中模型的在线辨识和控制器的在线设计问题,以达到对具有大时滞、不确定非线性特性的大滞后系统实现高精度输出跟踪控制。仿真结果表明:这一方法具有精确的跟踪控制效果。更多还原

【Abstract】 Aiming at the control feature of large delay system,a 2-rank derivative multi-step neural network predictive model and the algorithm were designed,and a fuzzy control scheme was presented based on the model with incomplete recurrent neural network and fuzzy CMAC controller.The problems of model online identification and controller online design in traditional adaptive control were solved,so that the high precision output fellow-up control of systems with large delay and uncertain nonlinear features can be realized.The simulation results show that the method has perfect control effect.更多还原