基于BP神经网络的飞行员着舰训练品质评估  

Pilot landing training quality evaluation based on BP neural network

作　　者：张海燕 闫文君[1] 张立民 李忠超 ZHANG Haiyan;YAN Wenjun;ZHANG Limin;LI Zhongchao(Naval Aviation University,Yantai 264001,China;Unit 31627 of Troop PLA,Shenzhen 518000,China;Unit 73022 of Troop PLA,Huizhou 516000,China)

机构地区：[1]海军航空大学,山东烟台264001 [2]中国人民解放军31627部队,广东深圳518000 [3]中国人民解放军73022部队,广东惠州516000

出　　处：《电子设计工程》2024年第9期37-41,共5页Electronic Design Engineering

基　　金：信息系统安全技术重点实验室基金(614211190404)。

摘　　要：舰载战斗机在深海作战中发挥着重要作用,飞行员着舰训练品质的高低直接影响着舰载机的战斗力,以往对飞行员着舰训练品质的评估采用人工方式,很少尝试神经网络。针对这方面的不足,提出了基于反向传播(BP)神经网络的飞行员着舰训练品质评估方法。利用着舰飞行参数、舰载机尾钩挂锁情况以及专家组评分,构建数据集;对网络进行训练和测试,确定网络参数,训练着舰评估网络模型;通过验证集对网络进行仿真验证,验证模型的可靠性。验证结果表明,该网络能较为准确地评估着舰分数和尾钩挂锁情况,可为飞行员着舰训练提供参考。Carrier-based fighter plays an important role in deep-sea combat,and the quality of pilot landing training directly affects the combat effectiveness of carrier-based fighter.In the past,the evaluation of pilot landing training was often conducted by artificial method,and neural network was seldom tried.Aiming at these shortcomings,a quality evaluation method of pilot landing training based on Back Propagation(BP)neural network is proposed.The data set is constructed by using the flight parameters of landing aircraft,the situation of tail hook and lock of carrier-borne aircraft and the score of expert group.Train and test the network,determine the network parameters,train and evaluate the network model;The reliability of the model is verified by the simulation of the network through the verification set.The verification results show that the network can accurately evaluate the landing score and the situation of tail-hook lock,which can provide a reference for pilot landing training.

关 键 词：BP神经网络 飞行员 着舰训练 品质评估 

分 类 号：TN02[电子电信—物理电子学]