一种基于期望模型的自适应Singer模型滤波算法

针对使用固定模型滤波算法跟踪机动目标时滤波精度依赖于模型固有参数的问题，提出了一种基于期望模型的自适应Singer模型滤波算法。首先利用3组代表不同机动强弱的典型Singer模型组成基础模型集合，然后通过实时计算目标综合残差确定目标机动等级，根据目标机动等级的变化来生成期望模型，并实时扩充基础模型集合进行交互式多模型（Interacting Multiple Model,IMM）滤波。该算法降低了对基础模型选取的依赖性，具有更好的环境适应性，在目标不同机动状态下都能进行准确跟踪。

An adaptive filtering algorithm for Singer model based on expectation model

In order to solve the problem that the filtering accuracy depends on the inherent parameters of the model when the fixed model filtering algorithm is used to track maneuvering targets,an adaptive Singer model filtering algorithm based on the expected model is proposed.At first,three groups of typical Singer models representing different maneuvers are used to form the basic model set.Then,the target comprehensive residual is calculated in real time to determine the maneuvering level of the target.The expected model is then generated according to the change of the maneuvering level of the target,and it is used to expand the basic model set in real time.The interactive multiple model(IMM) filtering is performed at each time based on the new extended model set.The proposed algorithm reduces the dependence on basic model,has better adaptability to the environment,and can accurately track targets in different maneuvering states.