基于变维交互式多模型算法的近区目标跟踪研究

雷达数据处理中的一个关键部分是目标跟踪,它负责对目标的点迹进行滤波,降低噪声影响.交互式多模型算法是一种基于广义伪贝叶斯算法的具有马尔科夫转移概率的多模型算法.它可以用多个模型匹配目标的不同机动模式,避免了单模型不能全程匹配带来的模型误差.近区目标往往具有较大的角速度和角加速度,模型集里包含角速度分量可以更好的匹配其机动形式.于是在交互式多模型的基础上提出了一种变维交互多模型算法,可以对不同维度的模型集进行交互滤波,以取得更好的跟踪精度.

Research of near zone target tracking method based on dimension changed IMM algorithm

Target tracking is a key part of radar data processing,which is responsible for the target point tracks' filtering.IMM is a multi-model algorithm based on generalized pseudo-Bayesian algorithm.IMM can use multiple models to match different maneuvering modes,avoiding the model error.Taking into account the dimensions may be different between the selected models,proposed a variable dimension IMM algorithm.The variable dimension IMM can interact with different models which have different dimensions to match the target better.