摘要
针对传统目标跟踪算法鲁棒性较差等问题,提出了一种基于当前统计模型的无迹卡尔曼滤波交互式多模型(IMM-CS-UKF)融合算法。在交互式多模型算法框架内,计算当前统计模型的概率,提高了统计模型的目标加速度和自适应性。该算法结合了交互式多模型和无迹卡尔曼滤波算法,具有对不同机动模式目标的自适应跟踪能力和精度高等优点。仿真结果表明,该算法对以多种机动策略实时机动的目标具有较好的跟踪性能。
Aiming at the problems of robustness weakness of traditional target tracking algorithm, an un- scented Kalman filter interactive multiple model based on current statistical model（ IMM -CS -UKF）fusion algorithm is proposed. In the interactive multiple model algorithm framework, the probability of current sta- tistical model is calculated, and the target acceleration and adaptability of statistical model are improved. The interactive multiple model unscented Kalman filtering algorithm is combined with this algorithm which has adaptive tracking ability for different maneuvering model target and has higher precision. The simula- tion results show that the algorithm has better tracking performance for multiple maneuvering model with re- al -time maneuvering target.
出处
《航天控制》
CSCD
北大核心
2015年第2期22-25,31,共5页
Aerospace Control
基金
国家自然基金(61171179
61227003
61301259)
山西省自然科学基金(2012021011-2)
高等学校博士学科点专项科研基金(博导类)(20121420110006)
山西省回国留学人员科研资助项目(2013-083)
关键词
无迹卡尔曼滤波
机动目标跟踪
交互式多模型
当前统计模型
Unscented Kalman filter
Maneuvering target tracking
Interactive multiple model
Current sta-tistical model
