摘要
提出一种改进的自适应新生目标GM-PHD算法。该算法以存活目标的量测更新权值构建“似然函数”,通过该函数确定量测来源并对新生目标权值做重分配,有效解决了归一化失衡问题。在量测方程高度非线性情况下,引入统计线性回归方法对量测方程进行线性化近似,求解新生目标预测均值和协方差。仿真结果表明,在新生目标信息先验缺失时,改进后的算法具有良好的跟踪精度和较低的计算量。
Proposing an improved adaptive nascent target GM-PHD algorithm.The algorithm constructs a“likelihood function”based on the updated weights of the surviving targets,through which it determines the source of the measurements and redistributes the weights of the newborn targets,thus effectively solving the problem of normalisation imbalance.At the same time,a statistical linear regression method is introduced to linearise the approximation of the measurement equations in the highly non-linear case of the measurement equations to achieve the solution of the predicted mean and covariance of the newborn target.Computer simulation results show that the improved algorithm has good tracking accuracy and low computational effort when the nascent target information is missing a priori.
作者
池桂林
胡磊力
周德召
CHI Guilin;HU Leili;ZHOU Dezhao(Science and Technologyon Electro-Optical Control Laboratory,Luoyang 471000,China;Luoyang Institute of Electro-Optical Equipment,AVIC,Luoyang 471000,China)
出处
《兵器装备工程学报》
CAS
CSCD
北大核心
2024年第S01期269-275,共7页
Journal of Ordnance Equipment Engineering
关键词
多目标跟踪
概率假设密度
自适应新生目标强度
随机有限集
multi-target tracking
probability hypothesis density filter
adaptive target birth intensity
random finite set
