摘要
针对传统的交互式多模型(IMM)算法通常采用相同维数的模型进行滤波,存在较大的模型误差以及当前统计模型(CS)中的参数需要合理设定的问题,提出一种变维自适应交互式多模型(AIMM)跟踪算法。该算法首先利用维数变换,将不同维数的模型转换为统一的维数进行交互滤波,使之适用于一般的机动目标,减少模型跟踪误差;然后通过引入由残差信息定义的调整因子对CS模型中的参数自适应调整,提高模型与实际运动模式的匹配程度;最后将参数调整后的CS模型反馈到变维IMM算法中,来改善跟踪性能。仿真实验表明,与传统变维IMM算法相比,文中所提算法在有效跟踪机动目标的同时,提高了目标的跟踪精度。
The traditional IMM filtering algorithm with model of the same dimension has the problem of a relatively large error, and it is difficult for Current Statistical （CS） model to set reasonable parameters. To solve the problems, we proposed a variable dimension adaptive IMM algorithm （AIMM） for tracking a maneuvering target. The algorithm firstly makes dimension variation processing to make the model with unified dimension, which is applicable to common maneuvering target, and can reduce the model tracking error. After that, the function based on residual information is used to adjust the CS model parameters, to improve the match degree between the model and the actual motion model. Finally, the improved CS mode is fed back to the variable dimension IMM for increasing tracking performance. The simulation results show that： Compared with the traditional variable dimension IMM algorithm, the proposed algorithm has higher tracking accuracy while keeping effective tracking on maneuvering target.
出处
《电光与控制》
北大核心
2015年第2期36-40,45,共6页
Electronics Optics & Control
基金
陕西省自然科学基金(2011JM8023)
关键词
机动目标跟踪
维数变换
交互式多模型算法
调整因子
maneuvering target tracking
variable dimension
interacting multiple model algorithm
adjustment factor
