基于压缩感知的PD雷达序贯扩展卡尔曼滤波跟踪方法

摘要

A new sequential extended Kalman filter (SEKF) based on compressive sensing (CS) is proposed to track ma neuvering targets for the pulse Doppler radar.The sparsity of target measurements in delay-Doppler plane is used to set up a sparse signal model in each pulse interval,and then Doppler measurements can be obtained through the reconstruction algorithm.Finally,SEKF is used to make filter update so as to attain the high-precision state estimation.In the process flow of tracking filter,CS processing is applied to improve estimate accuracy of delay and Doppler for targets,and SEKF is applied to reduce the nonlinearity between Doppler measurements and target motion state through adding the pseudo.Simulation results show that compared to the traditional SEKF method and the existing CS based tracking method,CS aided SEKF method has the highest tracking accuracy.Therefore,this proposed algorithm is validated to enhance the tracking performance of maneuvering target.%提出一种新的基于压缩感知(Compressive Sensing,CS)处理的序贯扩展卡尔曼滤波(Sequential ExtendedKalman Filter,SEKF)方法,以用于脉冲多普勒(Pulse Doppler,PD)雷达机动目标跟踪.利用目标在时延多普勒平面内的稀疏特点建立稀疏量测模型,然后通过压缩采样匹配重构方法获得目标的多普勒量测值,并用SEKF方法进行滤波更新,以改善目标状态的估计性能.在滤波过程中,应用CS处理可改善目标时延和多普勒的估计精度,而应用SEKF则可通过加入伪量测减小多普勒量测和目标运动状态之间的非线性误差.仿真实验结果显示,基于CS辅助SEKF方法的跟踪精度明显高于传统的SEKF方法以及已有基于CS的跟踪方法,所以本文所提出的方法对机动目标有更好的跟踪性能.