基于高斯混合伯努利滤波的超宽带传感器扩展目标跟踪

Extended Target Tracking of UWB Sensor Based onGauss Mixture Bernoulli Filtering

针对超宽带传感器扩展目标跟踪问题,提出了一种高斯混合伯努利滤波算法。算法基于梯度算子实现线性高斯目标和测量模型下的递归滤波,利用高斯混合模型获得目标的后验强度,提出一种基于层次划分密度的聚类优化来实现目标状态的提取,从而来估计扩展目标的时变散射点数。超宽带雷达传感器的仿真实验结果表明,所提算法可以在检测不确定度、目标测量率不确定度、噪声和假警报存在的情况下,能够有效联合检测和跟踪目标,在保证计算效率的同时,提高了跟踪的精确度及稳定性。

Aiming at the extended target tracking problem of ultra-wideband sensors,a Gaussian mixture Bernoulli filtering algorithm is proposed.First the linear Gaussian target and recursive filtering under the measurement model based on the gradient operators are and implemented by the algorithm then the Gaussian mixture model is used to obtain the posterior strength of the target,and finally a clustering optimization based on the hierarchical partition density is proposed to achieve the extraction of the target state so as to estimate the number of time-varying scattering points of the extended targets.The simulation experiment results of the ultra-wideband radar sensor show that the algorithm proposed in this paper can effectively detect and track the target in the presence of detection uncertainty,target measurement rate uncertainty,noise and false alarms,while calculation efficiency is being ensured,the accuracy and stability of tracking are improved.