摘要
针对当前机动目标跟踪领域中多模型算法存在的问题,提出一种基于粒子滤波的模型自适应机动目标跟踪算法.首先,依据前一时刻每个粒子采用的模型状态和模型间的状态转移概率,实现对当前时刻模型的采样;然后,将采样模型用于对当前粒子的预测,并根据当前时刻得到的量测数据实现对预测粒子权值的度量;最后,通过重采样策略和概率最大化原则完成对模型的合理选择和状态的有效估计.仿真实验验证了该算法的有效性.
Aiming at the existing problems of current multi-model algorithm for maneuvering target tracking, the paper proposes a model adaptive maneuvering target tracking algorithm based on particle filtering. Firstly, the algorithm achieves the sampling for current moment model according to the previous moment particle model information and the model transfer probability. Then, the prediction for current particle is accomplished by combining the model sampling result with state transfer equation, and the weight of prediction particle is measured based on current moment measurement. Finally, the re-sampling step and the principle of probability maximization are utilized to realize reasonable selection for model and effective estimation for state. Simulation results show the effectiveness of the algorithm.
出处
《控制与决策》
EI
CSCD
北大核心
2008年第12期1333-1337,共5页
Control and Decision
基金
国家自然科学基金重点项目(60634030)
国家自然科学基金项目(60702066)
中国航空工业第一集团公司航空基金项目(2007ZC53037)
中国航天科技集团公司航天科技创新基金项目(CASC0214)
关键词
机动目标跟踪
粒子滤波
交互式多模型
模型自适应
Maneuvering target tracking
Particle filtering
Interacting multiple model（IMM）
Model adaptive
