小波神经网络UCAV飞行控制系统  

A Study on Flight Control System Based on Wave-net

作　　者：杜海文[1] 翁兴伟[1] 赵辉[1] 傅裕松[1] 

机构地区：[1]空军工程大学工程学院,西安710038

出　　处：《火力与指挥控制》2012年第3期31-34,共4页Fire Control & Command Control

基　　金：国家"863计划"基金(2008AAJ205);军队重点科研基金(KJ06703);教研新星基金资助项目

摘　　要：针对采用H∞控制设计方法需求解难度很大的偏微分方程,而导致工程实现难度大的问题,提出了采用小波神经网络进行无人攻击机(UCAV)飞行控制系统设计的方法,该方法集小波变换和神经网络的优点于一身,简化了网络训练,避免了系统达到局部最优解,使得系统具有更好的逼近精度。通过仿真实验表明,利用小波神经网络进行UCAV控制律设计,使系统具有较好的跟踪性和鲁棒性,避免了精确模型的建立,便于分析系统的稳定性。To study the problem of using H∞ to control and design flight control system of UCAV is to extract difficult partial differential equation and which is difficult in engineering application,a way using wave-net to design flight control system of UCAV is proposed.The way has the merits of wavelet transform and neural networks,which simplifies the network training and avoids partial optimal value of the system and makes the system have better approach precision.The simulation results shows that the designed controll law of UCAV based on wave-net makes the system have good stability and robustness,which avoids building precise model and is convenient to analyze the stability of system.

关 键 词：无人攻击机 飞行控制系统 小波神经网络 控制律 非线性不确定系统 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]