| 引用本文: | 徐洋,方洋旺,伍友利,等.干扰及机动条件下的比例导引智能调控策略.[J].国防科技大学学报,2019,41(3):137-145.[点击复制] |
| XU Yang,FANG Yangwang,WU Youli,et al.Proportional guidance intelligent regulation strategy under the infrared interference and maneuvering[J].Journal of National University of Defense Technology,2019,41(3):137-145[点击复制] |
|
| |
|
|
| 本文已被：浏览 6626次 下载 5411次 |
| 干扰及机动条件下的比例导引智能调控策略 |
| 徐洋1, 方洋旺1, 伍友利1, 杨鹏飞2, 张丹旭1 |
|
(1 .空军工程大学 航空工程学院， 陕西 西安 710038;2. 军事科学院 评估论证研究中心， 北京 100091)
|
| 摘要: |
| 由于红外诱饵干扰样式复杂、目标机动形式多变导致传统比例导引律极易被干扰。为提高采用比例导引方法的导弹性能，提出一种利用径向基函数网络调控比例系数及导弹发射时机的智能导引律。以飞行时间及脱靶量为参考，通过构建加权型指标函数将求解最优比例系数及发射时机问题转化为单目标优化问题；引入量子粒子群算法求解最优决策参量，并以其作为网络输出，干扰样式作为网络输入，离线训练径向基函数网络；为提高训练效率，结合K-means及K最近邻算法初始化径向基函数网络。仿真结果表明，当存在红外诱饵干扰时，智能导引律性能优于扩展比例导引律及自适应滑模导引律。 |
| 关键词: 红外诱饵 量子粒子群 比例导引律 径向基网络 |
| DOI：10.11887/j.cn.201903021 |
| 投稿日期：2018-05-21 |
| 基金项目:博士后创新基金资助项目（BX201700104） |
|
| Proportional guidance intelligent regulation strategy under the infrared interference and maneuvering |
| XU Yang1, FANG Yangwang1, WU Youli1, YANG Pengfei2, ZHANG Danxu1 |
|
(1. Aeronautics Engineering College, Air Force Engineering University, Xi′an 710038, China;2. Research Center for Assessment and Argumentation, Academy of Science, Beijing 100091, China)
|
| Abstract: |
| Due to the complicated interference pattern of the infrared decoy and the variable shape of the target maneuver, the traditional proportional guidance law is easily interfered. In order to improve the performance of missiles using the proportional guidance method, an intelligent guidance law that uses the RBF(radial basis function) network to control the proportional coefficient and the timing of missile launch is proposed. Taking the flight time and miss distance as reference, the weighted index function is used to transform the optimal proportional coefficient and launch timing problem into a single objective optimization problem. The quantum particle swarm optimization algorithm is introduced to solve the optimal decision parameters which are used as the network output, and the interference pattern is used as the network input to train the RBF network offline. To improve the training efficiency, the RBF network is initialized by combining the K means and KNN(K nearest neighbors) algorithms. Simulation results show that the intelligent guidance law is better than the extended proportional guidance law and the adaptive sliding mode guidance law when there is infrared decoy interference. |
| Keywords: infrared decoy quantum particle swarm optimization proportional guidance radial basis function network |
|
|
