变结构多模型状态估计器方法适用于雷达跟踪空中机动目标。本文研究和设计了两种自适应时变模式集交互多模型(IMM)跟踪滤波器;切换网格(SG)IMM和自适应网格(AG)IMM算法。针对不同的飞行想定,通过Monte Carlo仿真方法对这些算法及相应的固定网格(FG)IMM滤波器的性能进行了评估和比较。仿真显示,对特定的机动目标跟踪问题,SGIMM和AGIMM跟踪滤波器在性能和计算量等方面要比固定结构多模型算法(FGIMM滤波器)有明显地提高。 Variable structure multi-model state estimator method is suitable for radar tracking aerial maneuvering target. In this paper, two kinds of adaptive time-varying (IMM) tracking filters, switched mesh (SG) IMM and adaptive mesh (AG) IMM algorithms are studied and designed. For different flight scenarios, the performance of these algorithms and the corresponding fixed-grid (FG) IMM filters were evaluated and compared using Monte Carlo simulations. The simulation shows that SGIMM and AGIMM tracking filters have higher performance and computational efficiency than the fixed-structure multi-model algorithm (FGIMM filter) for certain maneuvering target tracking problems.