多传感器粒子滤波融合跟踪算法  

Multi-sensor Fusion Tracking Algorithm Based on Particle Filter

作　　者：李龙[1] 秦超英[1] 

机构地区：[1]西北工业大学理学院应用数学系,西安710129

出　　处：《科学技术与工程》2010年第32期7947-7950,7955,共5页Science Technology and Engineering

摘　　要：对于非线性非高斯环境中的多传感器分布式状态估计问题,提出了一种基于二阶中心差分粒子滤波方法的融合跟踪算法。通过对量测方程的非线性分析,利用粒子滤波器计算目标状态估计值,以在线自适应加权融合算法的方式得到系统最优估计。仿真结果表明,与采用扩展卡尔曼滤波的方法相比,该算法具有更高的估计精度。In order to solve the distributed multi-sensor state estimation problem of non-Gaussian nonlinear system,a fusion tracking algorithm based on second-order central difference particle filter is proposed. It uses particle filter to calculate state estimated values by the nonlinear analysis of measurement equation,and then the system optimal estimation is obtained in the adaptive weighted fusion algorithm way. The simulation results show that compared with the extended Kalman filter,the proposed algorithm improves the estimation accuracy.

关 键 词：多传感器 分布式 粒子滤波 非线性非高斯 中心差分 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]