融合粒子滤波与交互多模算法的优势，提出了一种基于进化粒子滤波的交互多模算法(EPF-IMM)。该算法将遗传进化思想引入到传统的粒子滤波，在粒子迭代中采用遗传算法中的编码、交叉、变异等算子实现粒子的自适应进化且隐含重采样，从而改进其粒子退化现象。然后利用粒子滤波信息，在交互多模型中进行更新运算。既解决了IMM算法对非线性、非高斯环境的适应性问题，又解决了PF的无关联对应模型问题。与标准IMM算法进行高机动目标跟踪性能比较，试验仿真结果表明，EPF-IMM算法的跟踪精度高。

An Interactive Multiple Model(IMM) algorithm based on Evolution Particle Filtering(EPF) was proposed by combing the advantages of particle filter with that of IMM. The genetic evolution ideology was introduced into traditional particle filter to improve the phenomenon of particle degradation by using operators of coding, intersection, and variation to achieve adaptive evolution of particles. Then the particle filtering information was adopted to update operation in IMM.The algorithm solves the problem existed in IMMs adaption to nonlinear and non-Gaussian situation，and also the problem of PF that there is no relevant model.Simulation was made for comparing the high-maneuvering target tracking performance with that of the standard IMM algorithm，and the result showed that EP-IMM algorithm can achieve a higher accuracy