CPHD(Cardinalized Probability Hypothesis Density)滤波是一种杂波环境下可变目标数的多目标跟踪算法,该文针对算法中存在的目标漏检问题提出一种改进算法,该算法在高斯混合框架下实现贝叶斯递归,通过对各个高斯分量进行标记,对目标...CPHD(Cardinalized Probability Hypothesis Density)滤波是一种杂波环境下可变目标数的多目标跟踪算法,该文针对算法中存在的目标漏检问题提出一种改进算法,该算法在高斯混合框架下实现贝叶斯递归,通过对各个高斯分量进行标记,对目标进行航迹关联,在此基础上对修剪合并后各个高斯分量的权值进行两次分配。首先对超过检测门限的高斯分量权值进行分配,有效解决了目标漏检问题,然后基于一个目标只可能产生一个观测的事实进行第2次分配,改善了目标发生交叉时的算法性能。实验结果表明,所提方法在多目标状态估计和航迹维持方面均优于普通的CPHD算法。展开更多
标准的带势概率假设密度(cardinalized probability hypothesis density,CPHD)滤波器是一个有效的多目标跟踪算法,但是它假定新生目标的强度函数先验已知,因而无法应用于新生目标在场景中任意位置出现的环境。针对此问题,提出一种单步...标准的带势概率假设密度(cardinalized probability hypothesis density,CPHD)滤波器是一个有效的多目标跟踪算法,但是它假定新生目标的强度函数先验已知,因而无法应用于新生目标在场景中任意位置出现的环境。针对此问题,提出一种单步初始化的高斯混合CPHD滤波器。该滤波器利用位置上远离当前时刻估计状态的观测值单步初始化新生目标。此外,多普勒信息一方面被用来初始化新生目标的速度,另一方面在滤波器更新步骤中,多普勒速度和位置观测信息采用串行更新方法处理。仿真结果表明,所提算法在目标数的估计精度和优化子模式分配距离方面优于已有算法。展开更多
In multiple extended targets tracking, replacing traditional multiple measurements with a rectangular region of the nonzero volume in the state space inspired by the box-particle idea is exactly suitable to deal with ...In multiple extended targets tracking, replacing traditional multiple measurements with a rectangular region of the nonzero volume in the state space inspired by the box-particle idea is exactly suitable to deal with extended targets, without distinguishing the measurements originating from the true targets or clutter.Based on our recent work on extended box-particle probability hypothesis density(ET-BP-PHD) filter, we propose the extended labeled box-particle cardinalized probability hypothesis density(ET-LBP-CPHD) filter, which relaxes the Poisson assumptions of the extended target probability hypothesis density(PHD) filter in target numbers, and propagates not only the intensity function but also cardinality distribution. Moreover, it provides the identity of individual target by adding labels to box-particles. The proposed filter can improve the precision of estimating target number meanwhile achieve targets' tracks. The effectiveness and reliability of the proposed algorithm are verified by the simulation results.展开更多
文摘CPHD(Cardinalized Probability Hypothesis Density)滤波是一种杂波环境下可变目标数的多目标跟踪算法,该文针对算法中存在的目标漏检问题提出一种改进算法,该算法在高斯混合框架下实现贝叶斯递归,通过对各个高斯分量进行标记,对目标进行航迹关联,在此基础上对修剪合并后各个高斯分量的权值进行两次分配。首先对超过检测门限的高斯分量权值进行分配,有效解决了目标漏检问题,然后基于一个目标只可能产生一个观测的事实进行第2次分配,改善了目标发生交叉时的算法性能。实验结果表明,所提方法在多目标状态估计和航迹维持方面均优于普通的CPHD算法。
文摘标准的带势概率假设密度(cardinalized probability hypothesis density,CPHD)滤波器是一个有效的多目标跟踪算法,但是它假定新生目标的强度函数先验已知,因而无法应用于新生目标在场景中任意位置出现的环境。针对此问题,提出一种单步初始化的高斯混合CPHD滤波器。该滤波器利用位置上远离当前时刻估计状态的观测值单步初始化新生目标。此外,多普勒信息一方面被用来初始化新生目标的速度,另一方面在滤波器更新步骤中,多普勒速度和位置观测信息采用串行更新方法处理。仿真结果表明,所提算法在目标数的估计精度和优化子模式分配距离方面优于已有算法。
文摘In multiple extended targets tracking, replacing traditional multiple measurements with a rectangular region of the nonzero volume in the state space inspired by the box-particle idea is exactly suitable to deal with extended targets, without distinguishing the measurements originating from the true targets or clutter.Based on our recent work on extended box-particle probability hypothesis density(ET-BP-PHD) filter, we propose the extended labeled box-particle cardinalized probability hypothesis density(ET-LBP-CPHD) filter, which relaxes the Poisson assumptions of the extended target probability hypothesis density(PHD) filter in target numbers, and propagates not only the intensity function but also cardinality distribution. Moreover, it provides the identity of individual target by adding labels to box-particles. The proposed filter can improve the precision of estimating target number meanwhile achieve targets' tracks. The effectiveness and reliability of the proposed algorithm are verified by the simulation results.
