尺度因子自适应的UKF算法在目标跟踪中的应用  

The Application of UKF with Scale Adaptive in Object Tracking

作　　者：侯建华[1] 刘倩[1] 笪邦友[1] 马晓路[1] 

机构地区：[1]中南民族大学电子信息工程学院,武汉430074

出　　处：《中南民族大学学报（自然科学版）》2012年第2期85-89,共5页Journal of South-Central University for Nationalities：Natural Science Edition

基　　金：国家自然科学基金资助项目(61141010);武汉市科技供需对接计划项目(201051824575)

摘　　要：针对传统的无迹卡尔曼滤波(UKF)算法不能根据场景变化而自适应调整尺度因子α的问题,提出了一种改进算法,该算法利用UKF非线性近似的预测值与真实值之间的误差来调节α,并对采样策略进行了修正.将此方法应用于目标跟踪的仿真实验表明:该算法与使用尺度因子最优经验值的UKF算法精度相当,具有很好的跟踪性能和实用性.To improve the adaptivity of the standard Unscented Kalman filter （UKF） to scene change, an adaptive UKF is proposed based on the scale factor. The scale factoris adjusted by the error between the non-linear approximation of the UKF prediction and the true value. Then the sampling strategy is also revised. The proposed method is applied to target tracking. The simulation results show that compared the standard UKF with the optimum empirical value, the new algorithm has the comparable tracking performance and satisfactory practicality in the application of object tracking.

关 键 词：无迹卡尔曼滤波 尺度因子自适应 比例修正 目标跟踪 实用性 

分 类 号：TP919.81[自动化与计算机技术]