摘要
针对临近空间高超声速滑翔飞行器机动模式复杂,单一运动学模型难以完成三维跟踪的问题,提出一种三维跟踪方法。将飞行器机动弹道分为纵向和横向弹道,根据飞行器机动特性,在纵向上将加速度建模为零均值的二阶时间自相关随机过程,在横向上采用Singer模型和匀加速模型进行交互多模型(interactive multiple model,IMM)滤波,引入无偏量测转换将球坐标系下的雷达观测模型转换为笛卡尔坐标系,避免了模型的非线性滤波。最后,在现有卡尔曼滤波基础上设计了一种基于多重渐消因子的自适应卡尔曼滤波方法,增强了模型对强机动的自适应跟踪能力。仿真实验表明,该算法在对高超声速滑翔飞行器进行三维跟踪时,能保持较好的稳定性和较高的跟踪精度。
In order to solve the problem that the maneuvering mode of hypersonic gliding vehicle is complex and the single kinematic model is difficult to complete the three-dimensional tracking,a three-dimensional tracking method is proposed.The maneuvering trajectory of the vehicle is divided into longitudinal and lateral trajectory.According to the maneuvering characteristics of the vehicle,the acceleration is modeled as a second-order time autocorrelation stochastic process with zero mean in the longitudinal direction,and the Singer+constant acceleration interactive multiple model(IMM)is used in the lateral direction.The unbiased measurement transformation is introduced to transform the radar observation model in spherical coordinate system to Cartesian coordinate system,which avoids the nonlinear filtering of the model.Finally,based on the existing Kalman filtering,an adaptive Kalman filtering algorithm based on multiple fading factors is designed to enhance the adaptive tracking ability of the model for strong maneuvers.The simulation results show that the algorithm can keep better stability and tracking accuracy when tracking hypersonic gliding target.
作者
张君彪
熊家军
兰旭辉
李凡
刘文俭
席秋实
ZHANG Junbiao;XIONG Jiajun;LAN Xuhui;LI Fan;LIU Wenjian;XI Qiushi(Department of Graduate, Air Force Early Warning Academy, Wuhan 430019, China;The Fourth Department, Air Force Early Warning Academy, Wuhan 430019, China;Unit 95980 of the PLA, Xiangyang 441000, China)
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2022年第2期628-636,共9页
Systems Engineering and Electronics
基金
军队重点科研课题(KJ20191A020148)
军事类研究生(JY2019B138)资助课题。
关键词
临近空间
高超声速飞行器
三维跟踪
滤波算法
交互多模型
near space
hypersonic vehicle
three dinensional(3D)tracking
filtering algorithm
interactive multiple model(IMM)
