基于Input Estimation的VSIMM机动目标跟踪

Input Estimation based Variable Structure Interactive Multiple Model for Maneuvering Target Tracking

IMM算法的跟踪性能很大程度上取决于模型集的选择。提出了基于Input Estimation的自适应改变模型集的变结构多模型算法IE_VSIMM。对IMM算法输出的状态估计及其误差协方差进行准Kalman滤波,由Input Estimation算法得到的加速度增量估计,可检测目标机动和生成新的模型集。修正过程则由IMM算法在新模型集上对状态估计及其误差协方差进行更正。仿真结果表明IE_VSIMM算法的跟踪性能更好。

Tracking performance of IMM depends greatly on the selection of model set..4 new method was proposed to determine adaptively the model set for IMM based on lnput Estimation （IE） method The acceleration increment could be estimated based on IE. When maneuver was detected, new model set could be determined Correction procedure would modify the state estimation and its error covariance with new model set. Then IMM estimation could go on with the newly determined model set. The presented IE_VSIMM has simpler algorithm structure and needs fewer rules and parameters. Simulation results show that IE_VSIMM has better tracking performance and can closely keep up with the acceleration change of maneuvering target.