计算机工程与应用
計算機工程與應用
계산궤공정여응용
COMPUTER ENGINEERING AND APPLICATIONS
2014年
10期
238-243,248
,共7页
隐性交叉遗传算法%广义回归神经网络%实时动态%导弹追踪目标%导航常量
隱性交扠遺傳算法%廣義迴歸神經網絡%實時動態%導彈追蹤目標%導航常量
은성교차유전산법%엄의회귀신경망락%실시동태%도탄추종목표%도항상량
recessive crossover genetic algorithm%general regression neural network%real time dynamics%missile tracking target%navigation constant
针对三维环境中导弹追踪目标时制导和控制算法复杂而导致计算量非常大的问题,提出了一种基于隐性交叉遗传算法优化广义回归神经网络的实时动态目标追踪模型。通过将导弹防御区离散化为多个小模块生成输入数据,并针对每个可接受的目标参数数据集,使用RCGA估算导航常量和导弹注意时间;利用输入和输出的目标参数集生成GRNN所需的训练数据集;针对任意位置的目标轨道,将训练后的GRNN应用于实时导弹导引系统的实现中。通过战术目标仿真模型验证了所提算法的有效性及可靠性,仿真结果表明,相比其他几种目标追踪算法,算法取得了更好的实时性和更高的目标定位精度,脱靶率接近零。
針對三維環境中導彈追蹤目標時製導和控製算法複雜而導緻計算量非常大的問題,提齣瞭一種基于隱性交扠遺傳算法優化廣義迴歸神經網絡的實時動態目標追蹤模型。通過將導彈防禦區離散化為多箇小模塊生成輸入數據,併針對每箇可接受的目標參數數據集,使用RCGA估算導航常量和導彈註意時間;利用輸入和輸齣的目標參數集生成GRNN所需的訓練數據集;針對任意位置的目標軌道,將訓練後的GRNN應用于實時導彈導引繫統的實現中。通過戰術目標倣真模型驗證瞭所提算法的有效性及可靠性,倣真結果錶明,相比其他幾種目標追蹤算法,算法取得瞭更好的實時性和更高的目標定位精度,脫靶率接近零。
침대삼유배경중도탄추종목표시제도화공제산법복잡이도치계산량비상대적문제,제출료일충기우은성교차유전산법우화엄의회귀신경망락적실시동태목표추종모형。통과장도탄방어구리산화위다개소모괴생성수입수거,병침대매개가접수적목표삼수수거집,사용RCGA고산도항상량화도탄주의시간;이용수입화수출적목표삼수집생성GRNN소수적훈련수거집;침대임의위치적목표궤도,장훈련후적GRNN응용우실시도탄도인계통적실현중。통과전술목표방진모형험증료소제산법적유효성급가고성,방진결과표명,상비기타궤충목표추종산법,산법취득료경호적실시성화경고적목표정위정도,탈파솔접근령。
Guidance and control algorithm is complex when missile tracking targets in three-dimensional environment, which will cause complex computing, for which a real-time dynamic tracking target model based on general regression neural network optimized by recessive crossover genetic algorithm is proposed. Missile defense is discretized into many little models so as to generating imputing data, RCGA is used to estimate the navigation constant and notice time for each acceptable target parameter data set. Inputted and outputted target parameter data sets are used to generate training set needed by GRNN. The trained GRNN is applied into implementation of missile guidance real-time system. The effectiveness and reliability of proposed algorithm has been verified by tactical target simulation model. Simulation results show that proposed algorithm has better instantaneity and higher target positional accuracy than several target tracking algorithms.