基于Unscented变换的强跟踪滤波器  

Strong tracking filter based on unscented transformation

作　　者：王小旭[1,2] 赵琳[1] 夏全喜[1] 郝勇[1] 

机构地区：[1]哈尔滨工程大学自动化学院,哈尔滨150001 [2]西北工业大学自动化学院,西安710072

出　　处：《控制与决策》2010年第7期1063-1068,共6页Control and Decision

基　　金：国家自然科学基金项目(60974104)

摘　　要：针对强跟踪滤波器(STF)存在要求非线性函数连续可微和强非线性时滤波性能不佳等理论局限性,提出一种基于Unscented变换的强跟踪滤波器(UTSTF).UTSTF基于STF的理论框架,采用Unscented变换代替STF中的雅可比矩阵计算;UTSTF兼具STF鲁棒性强、Unscented变换精度高、实现简单的优点,有效克服了STF的理论局限性.仿真实例验证了UTSTF的有效性.Strong tracking filter （STF） has some theoretical limitations including bad filtering performance in strong nonlinear and calling for nonlinear function continuous and differentiable. Therefore,STF based on unscented transformation （UTSTF） is proposed. UTSTF views STF as the basic theory framework and makes unscented transformation take place of calculating nonlinear function Jacobian matrix,so it combines strong robustness of STF with high accuracy and easy implementation of unscented transformation. The simulation examples show the effectiveness of UTSTF.

关 键 词：非线性 强跟踪滤波器 Unscented变换 鲁棒性强 精度高 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]