改进的多模型粒子滤波弱小目标检测前跟踪方法

Improved Multiple Model Particle Filter Track-before-detect Algorithm for Maneuvering Weak Target

检测前跟踪是解决目标信噪比(SNR)较低的情况下目标检测与跟踪的有效方法。目前主要的研究方法有多模型粒子滤波弱小目标检测前跟踪方法(MMPF-TBD),该方法在目标出现较强的机动时,目标的检测性能会严重下降甚至出现漏检。该文针对该问题提出了一种改进的多模型粒子滤波弱小目标检测前跟踪方法(IMMPF-TBD),该方法可以降低模型之间转移计算复杂度,并且有效地提高模型的使用效率和目标的检测性能。仿真实验结果表明相比于MMPF-TBD,IMMPF-TBD能够有效地提高机动目标的检测性能。

Tarck-before-detect is an effective method to solve the target detection and tracking under the conditions of low SNR.At present,the main research method for maneuvering weak target is multiple model particle filter track-before-detect algorithm MMPF-TBD.In this way,when the target has a strong maneuver,the detection performance of the target will be seriously degraded or even missed.Therefore,this paper proposes an improved multi-model particle filter track-before-detect IMMPF-TBD algorithm.This method can solve the transfer computational complexity between models and effectively improve the efficiency of the model.This method can reduce the transfer computational complexity between models and effectively improve the efficiency of the model and the detection performance of the target.Simulation results show that IMPM-TBD can effectively improve the detection performance of maneuvering targets compared to MMPF-TBD.