群目标跟踪自适应IMM算法

发布时间：2018-01-10 13:40

本文关键词：群目标跟踪自适应IMM算法　出处：《哈尔滨工业大学学报》2016年10期 　论文类型：期刊论文

【摘要】：为提高对机动群目标在高量测误差下的跟踪性能,提出了一种自适应IMM群目标跟踪算法.首先,在群质心状态估计中,引入带有多重次优渐消因子的强跟踪滤波算法,提高机动阶段时对群质心状态估计的精度.其次,在扩展状态估计中,考虑量测精度对于扩展状态的影响,将量测误差和扩展状态同时纳入到量测似然函数的构建中,应用新息计算和渐消记忆迭代过程自适应更新量测误差协方差矩阵.最后,通过quasi-Bayesian方法自适应更新模型转换概率,利用量测数据修正模型转换概率,抑制非匹配模型作用,放大匹配模型作用,实时匹配跟踪模型与目标运动状态.仿真实验结果表明,该方法有效提高了对群质心状态和扩展状态的估计精度.
[Abstract]:In order to improve the tracking performance of maneuvering group targets with high measurement error, an adaptive IMM group target tracking algorithm is proposed. A strong tracking filtering algorithm with multiple suboptimal fading factors is introduced to improve the accuracy of mass center state estimation in maneuvering phase. Secondly, the effect of measurement precision on extended state is considered in extended state estimation. The measurement error and the extended state are incorporated into the construction of the measurement likelihood function at the same time, and the measurement error covariance matrix is updated adaptively by using the innovation calculation and the iterative process of fading memory. Finally. The model transformation probability is updated adaptively by quasi-Bayesian method, and the model transformation probability is corrected by measuring data to suppress the effect of non-matching model and amplify the function of matching model. The simulation results show that the proposed method can effectively improve the estimation accuracy of group centroid state and extended state.
【作者单位】：空军工程大学防空反导学院;
【基金】：国家自然科学基金(61501495)
【分类号】：TN713
【正文快照】： 群目标跟踪场景,如机群编队和地面移动车队,一系列拥有类似运动方式的空间临近目标组成跟踪对象.在群目标跟踪问题中,群目标的量测数目是时变的,传统多目标跟踪算法难以实现对群目标的稳定跟踪[1].文献[2]按照群内目标数目将群目标跟踪分成两类:大群目标跟踪和小群目标跟踪.其

【相似文献】