无人机涡喷发动机的BP网络控制

该文介绍了由反向传播神经网络（BP网络）构成神经网络PID控制系统的基本结构和原理,针对传统BP算法存在的收敛速度慢、容易陷入局部极小值的缺陷,选用了标准的数值优化算法（LM算法）,此算法具有极好的快速性;对某无人机涡喷发动机的神经网络PID控制进行了仿真研究,并与常规PID控制进行了比较研究,仿真结果表明：采用此算法构成的神经网络PID控制具有响应速度快、稳态误差小、算法简单的优点;用一个神经网络作为控制器,控制地面和空中多点对象是可行的,但要进行实际应用,还有待于进一步的研究.

The PID Neural Network Control of The Pilotless Turbojet Engine

This thesis introduces the structure and principle of the PID control system comprised of the BP neural network.Instead of the BP arithmetic which has certain deficiency of conducting the simulation slowly and being vulnerable to the minimum value,this thesis adopts the LM arithmetic which is fast in conducting simulation.By conducting simulation to the PID neural network control of the pilotless turbojet engine and comparing the simulating results with those by the conventional PID control method,conclusion could be safely drawn that the PID controller comprised of BP neural network is proved quicker in response,lesser in stable error and simpler in arithmetic method;and that it is theoretically possible to control several objects in the air and on the ground with the neural network as the controller,which still needs further study in order to be put in practise.