基于RB粒子滤波的多传感器目标跟踪融合算法

构建面向多传感器信息融合系统的粒子滤波(PF)器是拓展采样型非线性滤波应用领域的关键,针对PF在多传感器融合目标跟踪系统的有效实现问题,提出了一种基于Rao-Blackwellized(RB)PF(RB-PF)的多传感器目标融合跟踪(MT-RB-PF)算法。首先,利用RB建模技术实现跟踪系统非线性状态估计的降维处理;其次,结合多传感器融合系统特点,给出一种多量测下粒子权重优化新方法用以改善粒子权重度量的可靠性和稳定性;最终,通过标准PF和卡尔曼滤波(KF)实现非线性和线性状态分量的估计,并利用状态重构方法构建当前时刻的状态估计值。理论分析和仿真实验验证了算法的有效性。

Multi-sensor target tracking fusion algorithm based on Rao-Blackwellised particle filter

The structure of particle filter for multi-sensor information fusion system is the key to expanding the application domain of sampling nonlinear filters.Aiming at the effective realization of particle filter in multi-sensor fusion tracking system,a novel multi-sensor fusion target tracking algorithm based on Rao-Blackwellised particle filter is proposed.In the new algorithm,the reduction of tracking system state dimension is firstly realized by the Rao-Blackwellised modeling technology.Secondly,combining with the characteristics of multi-sensor fusion system,a new weight optimization method is used to improve the reliability and stability of particle weight.Finally,the system nonlinear and linear state components are respectively estimated by particle filter and Kalman filter,and the system state estimation is achieved by the state reconstruction method at the current time.The theoretical analysis and experimental results show the feasibility and efficiency of the proposed algorithm.