一种神经形态马赫数PID自适应控制器研究  

Research of Self-Adaptive PID Mach Controller Based on Neural Network

作　　者：赵书军[1] 施洪昌[1] 唐晓东[1] 

机构地区：[1]中国空气动力研究与发展中心,四川绵阳621006

出　　处：《计算技术与自动化》2004年第1期100-104,共5页Computing Technology and Automation

摘　　要：本文从自适应控制和优化学习算法的角度出发,将传统的PID控制和神经网络BP算法相结合,构造了具有自适应、自学习功能的神经形态PID自适应马赫数控制器(NNCPID)。通过对网络控制模型的训练、学习和系统仿真,验证了该NNCPID控制器在改善马赫数控制性能、提高系统实时性、鲁棒性等方面,具有很大的优越性。From the point of self-adaptive control and optimizing study algorithm, this thesis combines traditional PID control and BP neural network algorithm together to construct an self-adaptive neural network Mach PID controller which is called NNCPID. The topology structure of NN is constructed, the NN model is trained and a series of simulation experiment are done. It is proved that NNCPID is capable of improving the dynamic, real-time and robust character of Mach controller.

关 键 词：马赫数控制 PID算法 神经网络 自适应控制器 飞行器 智能控制 NNCPID控制器 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]