一种简化的拟蒙特卡洛-高斯粒子滤波算法
发布时间：2019-05-23 12:29
【摘要】：提出了一种简化的拟蒙特卡洛-高斯粒子滤波(QMC-GPF)算法(SQMC-GPF),以解决将QMC方法应用于GPF时计算复杂度高、运算量大的问题。该算法中,在连续的迭代滤波过程开始之前,首先利用QMC采样产生单位拟高斯分布粒子集,然后用其线性变换产生GPF算法中需要的高斯分布粒子集,省去了重新进行QMC采样步骤。该算法简化了新粒子集的产生过程,减少了运算量和滤波时间,增强了算法的实时性。将粒子滤波算法(PF)、GPF算法、QMC-GPF算法和SQMCGPF算法用于单变量非静态增长模型(UNGM)和二维纯角度跟踪模型(BOT)的仿真结果表明,SQMC-GPF算法的滤波性能与QMC-GPF算法的滤波性能相近,但有更为明显的速度优势,具有重要的实际应用价值。
[Abstract]:A simplified quasi-Monte Carlo Gao Si particle filter (SQMC-GPF) algorithm is proposed to solve the problem of high computational complexity and large computational complexity when the QMC method is applied to GPF. In this algorithm, before the continuous iterative filtering process begins, the unit quasi-Gaussian distribution particle subset is generated by QMC sampling, and then the Gao Si distribution particle subset needed in GPF algorithm is generated by its linear transformation. The QMC sampling step is omitted. The algorithm simplifies the generation process of new particle subset, reduces the amount of computation and filtering time, and enhances the real-time performance of the algorithm. The particle filter algorithm (PF), GPF algorithm, QMC-GPF algorithm and SQMCGPF algorithm are applied to the simulation results of single variable non-static growth model (UNGM) and two-dimensional pure angle tracking model (BOT). The filtering performance of SQMC-GPF algorithm is similar to that of QMC-GPF algorithm, but it has more obvious speed advantages and has important practical application value.
【作者单位】： 江苏科技大学电子信息学院;
【基金】：国家自然科学基金资助项目(61401179)
【分类号】：TN713
本文编号：2483895
[Abstract]:A simplified quasi-Monte Carlo Gao Si particle filter (SQMC-GPF) algorithm is proposed to solve the problem of high computational complexity and large computational complexity when the QMC method is applied to GPF. In this algorithm, before the continuous iterative filtering process begins, the unit quasi-Gaussian distribution particle subset is generated by QMC sampling, and then the Gao Si distribution particle subset needed in GPF algorithm is generated by its linear transformation. The QMC sampling step is omitted. The algorithm simplifies the generation process of new particle subset, reduces the amount of computation and filtering time, and enhances the real-time performance of the algorithm. The particle filter algorithm (PF), GPF algorithm, QMC-GPF algorithm and SQMCGPF algorithm are applied to the simulation results of single variable non-static growth model (UNGM) and two-dimensional pure angle tracking model (BOT). The filtering performance of SQMC-GPF algorithm is similar to that of QMC-GPF algorithm, but it has more obvious speed advantages and has important practical application value.
【作者单位】： 江苏科技大学电子信息学院;
【基金】：国家自然科学基金资助项目(61401179)
【分类号】：TN713
【相似文献】
相关期刊论文 前1条
1 万某峰;赵长胜;;UKF滤波中蒙特卡洛采样策略比较分析[J];测绘通报;2012年12期
，本文编号：2483895
本文链接：https://www.wllwen.com/kejilunwen/dianzigongchenglunwen/2483895.html
