一种改进的高超声速滑翔目标跟踪方法

An improved hypersonic glide target tracking method

针对临近空间高超声速滑翔目标跟踪问题,提出一种基于反向传播神经网络修正改进迭代扩展卡尔曼滤波(Back Propagation Neural Network-aided Improved Iterative Extended Kalman Filter, BP-IIEKF)的目标轨迹跟踪方法。在雷达站坐标系下建立目标运动模型和量测模型。引入阻尼因子修正IEKF算法中的协方差预测矩阵,并定义算法的代价函数,给出迭代终止条件,保证了算法收敛精度,减小状态的观测更新误差,提高了目标状态估计精度。利用BP神经网络修正滤波结果,补偿系统滤波误差,进一步提高了跟踪精度。仿真结果表明所提算法对高超声速滑翔目标具有更高的跟踪精度。

For the problem of tracking near space hypersonic glide target(HGT),a method of tracking HGT was proposed based on the back propagation neural network-aided improved iterative extended Kalman filter.The target motion model and the measurement model were established in the radar coordinate.The damping factor was introduced to modify the covariance prediction matrix in IEKF algorithm,the cost function of the algorithm was defined,and the iteration termination condi-tion was given.The convergence accuracy of the algorithm was guaranteed,the observation update error of the state was reduced,and the target state estimation accuracy was improved.The BP neural network was employed to correct the filtering results to compensate for the filtering errors of the entire system modeling,and to improve the tracking accuracy.Simulation results show that the proposed algorithm has a higher tracking accuracy for HGT.