基于自适应IMM算法的高超声速飞行器轨迹预测

Trajectory Predicting of Hypersonic Vehicle Using Adaptive Interactive Multiple Model Filter

针对非弹道式高超声速飞行器的周期跳跃运动特点,提出了一种基于多种运动模型的自适应交互多模型(IMM)方法对目标运动状态进行估计。建立了目标的常速度、常加速度、Singer和蛇形机动模型,以及雷达的量测模型。自适应IMM算法包括混合概率计算、输入融合(混合交互)、模型条件滤波、模型概率更新和滤波估计融合等步骤。算法融合了Singer和蛇形机动模型,能较好地适应有幅值变化和频率衰减的模型状态估计,并用滤波估值和双正弦和函数拟合加速度,进一步递推目标运动轨迹。仿真结果表明:自适应IMM算法对高超声速目标有较好的弹道跟踪和预测精度。

Aim to the characteristic of cyclic jump motion of non-ballistic hypersonic target,an adoptive interactive multiple model（IMM）algorithm based on multiple motion modes was put forward to estimate the target motion in this paper.The target modes of constant speed,constant acceleration,Singer and snake maneuvering and measurement mode were established.The adoptive IMM algorithm included five steps which were mixing probability computation,input fusing（mixing interacting）,mode conditional filtering,mode probability renew and filtering estimation fusing.The Singer and snake maneuvering mode were included in the algorithm,which could be suitable to the situation with amplitude changing and frequency attenuation finely.The acceleration fitting was made by filtering estimation and dual-sinusoid sum function to propagate the target trajectory.The simulation results showed that the adoptive IMM algorithm had good performance for trajectory tracking and prediction.