基于粒子滤波的机动目标跟踪改进算法  

Improved algorithm of maneuvering target tracking based on particle filter

作　　者：张苗辉[1] 辛明[2] 刘先省[1] 

机构地区：[1]河南大学先进控制与智能信息处理研究所,河南开封475001 [2]河南大学计算机与信息工程学院,河南开封475001

出　　处：《系统工程与电子技术》2008年第5期949-951,共3页Systems Engineering and Electronics

基　　金：河南省高校杰出科研人才创新工程项目(2003KYCX003);河南省高校创新人才培养工程;河南省自然科学基金(0411010400)资助课题

摘　　要：粒子滤波是指利用Monte Carlo仿真方法处理递推估计问题的非线性滤波算法,这种方法不受模型线性和Gauss假设的约束,是一种处理非线性非高斯动态系统状态估计的有效算法。在粒子滤波的基础上融合扩展卡尔曼滤波(EKF)算法,融合后的新算法在计算提议概率密度分布时,粒子的产生充分考虑当前时刻的量测,使得粒子的分布更加接近状态的后验概率分布。仿真结果表明,该算法对机动目标有更好的跟踪效果。The particle filter is an effective algorithm for the state recursive estimation in nonlinear and non Gaussian dynamic systems by utilizing Monte Cario simulation. Under the theory framework of particle filter, a new algorithm is presented, which combines the particle filter algorithm with extend Kalman filter algorithm. When the new algorithm calculates the proposed probability density distribution, the sampling particles can uti lize the system current measures. That gets the particles distribution more approach to the station posterior distribution. The simulation results show that the proposed algorithm can effectively track maneuvering target with much better performance.

关 键 词：机动目标跟踪 粒子滤波 扩展卡尔曼滤波 非线性非高斯 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]