基于UKF的曲线模型自适应跟踪算法  

Adaptive Tracking Algorithm of Curvilinear Model Based on Unscented Kalman Filter

作　　者：胡傲[1] 冯新喜[1] 李鸿艳[1] 李阳[1] 

机构地区：[1]空军工程大学电讯工程学院,陕西西安710077

出　　处：《探测与控制学报》2010年第2期73-77,82,共6页Journal of Detection & Control

基　　金：国家自然科学基金项目资助(60774091);陕西省自然科学基金项目资助(2007-24)

摘　　要：针对传统曲线跟踪模型中的切向加速度不能自适应调节这一缺点,在标准曲线模型的基础上,提出了一种新的自适应跟踪算法。将转弯角速率和切向加速度都看作是目标的状态变量,用不敏卡尔曼滤波算法对扩维后的状态变量进行估计。这种处理方式不仅较好地解决了原来算法中存在的强非线性问题,同时也增强了算法的鲁棒性。理论分析和仿真实验都表明,该算法适应性较强,跟踪精度较高,可以直接应用于工程实践。For the problem that the traditional curvilinear tracking model can not adapt its tangential acceleration in target tracking,a new algorithm based on normal curvilinear model is developed, which takes the state vector with turn rate and tangential acceleration as state vectors additionally and then estimates the target states with unscented Kalman filter Algorithm. By this means, the new algorithm does not only solve the problems of strong nonlinearity in the traditional algorithm,but also strengthens its robustness. Theoretical analysis and simulation results also show that the proposed algorithm can be directly applied to engineering for it is considerably self-adaptive and performs great accuracy in target tracking.

关 键 词：雷达数据处理 机动目标跟踪 曲线跟踪模型 不敏卡尔曼滤波 

分 类 号：TN953[电子电信—信号与信息处理] TP391.9[电子电信—信息与通信工程]