摘要
针对高斯混合概率假设密度(Gaussian mixture probability hypothesis density,GMPHD)滤波算法中的机动目标跟踪问题,提出了一种修正马尔可夫(Markov)矩阵的IMM-GMPHD算法.算法对交互多模型(interacting multiple model,IMM)作出修正,通过模型概率的变化率来修正Markov矩阵,增大了匹配模型在交互中的权重,抑制了非匹配模型的影响,提高了算法的估计精度;利用GMPHD滤波器作为IMM算法的滤波实现方式,完成了对多机动目标的数目和状态估计.仿真实验表明,所提算法能快速匹配目标模型的变化,提高匹配模型概率,改善滤波器性能.
In order to track maneuvering targets with Gaussian mixture probability hypothesis density （GMPHD） filtering algorithm, an IMM-GMPHD algorithm with Correctional Markov Matrix is proposed. The proposed algorithm amends the Markov matrix in interacting multiple model （IMM） by the rate of change in model probability~ this method can increase the weight of matching model in the interaction, restrain the bad influence of non-matching model and improve the algorithm estimation precision. The algorithm utilizes the GMPHD filter as the implementation of IMM to obtain the estimation of multiple maneuvering target state and the number. Simulation experiments show that the proposed algorithm can quickly match the target model, enhance the matching model probability and improve the performance of the filter.
出处
《军械工程学院学报》
2016年第1期40-45,共6页
Journal of Ordnance Engineering College
关键词
概率假设密度
高斯混合
机动目标
交互多模型
马尔可夫矩阵
probability hypothesis density
Gaussian mixture
maneuvering targets
interactingmultiple model
Markov matrix
