Gaussian mixture particle Cardinalized probability hypothesis density based passive bearings-only multi-target tracking
高斯混合粒子Cardinalized概率假设密度滤波被动测角多目标跟踪

When the number of targets is unknown or varies with time, multi-target state and measurements are expressed as random sets and the multi-target tracking problem is addressed by the Cardinalized probability hypothesis density(CPHD) filter, which propagates not only the probability hypothesis density(PHD) of the joint distribution but also the full probability distribution on target number. However, the CPHD can not provide a closed-form solution to the nonlinear problem occurred in the passive bearings-only multi-target tracking system. A novel Gaussian mixture particle CPHD(GMPCPHD) filter is presented in the paper. The PHD is approximated by a mixture of Gaussians, which avoids clustering in the determination of target states. In addition, Quasi-Monte Carlo integration method is introduced to approximate the prediction and update distributions of target states. Simulation results verify the effectiveness of the proposed GMPCPHD.