模糊小波神经网络控制器在隧进殉爆控制系统中的应用  

Application of Fuzzy Wavelet Neural Network Controller in Chiseling Detonation Control System

作　　者：蔡磊[1] 戴革林[2] 陆廷金[2] 袁冬根[1] 

机构地区：[1]空军工程大学工程学院,西安710038 [2]徐州空军学院航空弹药系,江苏徐州221000

出　　处：《电光与控制》2009年第1期59-62,66,共5页Electronics Optics & Control

基　　金：军队科研基金支持项目

摘　　要：采用模糊小波神经网络控制器对隧进殉爆系统进行控制,快速、准确、安全地排除机场未爆弹。模糊小波神经网络控制器采用小波函数作为模糊隶属函数,将模糊控制与神经网络相结合进行模糊推理。利用遗传算法全局搜索的特点,对神经网络参数进行离线优化;然后利用参数自调整的自主适应环境变化能力,对网络参数进行在线调整。实验结果表明:设计的控制器具有良好的鲁棒性,可以保证隧进殉爆系统在未爆弹弹道中顺利前行,完成未爆弹排除任务。The chiseling sympathetic detonation system is a nonlinear complicated system, and may be influenced by many factors during the process of Airport Unexploded Ordnances （UXO） detonation. Fuzzy wavelet neural network controller is used to control the chiseling detonation system for excluding UXOs rapidly and safely. It uses wavelet basis function as membership function, and combines fuzzy control and neural network together for fuzzy reasoning. Firstly, global search of genetic algorithm is used to optimize the fuzzy neural network＇s parameters off-line. Then, the self-adaptability of parameters to environment is used for adjusting the network parameters on-line. The results of the test proved that this controller has good robustness, which can guarantee the chiseling detonation system to move along the trajectory of UXO, and can fulfill the task of UXO obviation.

关 键 词：模糊小波神经网络 隧进殉爆系统 未爆弹排除 机器人 机场 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计]