任意相关噪声线性系统异步状态向量融合

基于线性无偏最小方差估计理论,提出了一种任意相关噪声线性系统异步状态向量融合算法.该算法将融合中心的采样周期设定为传感器测量周期的最小公倍数,使得传输到融合中心的局部状态估计在每个周期内具有相同的数目,减少了跟踪滤波的计算量.在跟踪滤波器的增益阵中引入测量噪声与过程噪声的相关量和测量噪声之间的相关量,增加了描述多传感器融合系统的信息量.仿真结果表明,状态向量融合算法比噪声不相关融合算法具有更好的跟踪性能,航迹跟踪的精度得到了改善.

Asynchronous State Vector Fusion of Linear Systems with Arbitrary Correlated Noises

Based on the linear unbiased minimum variance estimation theory, an asynchronous fusion algorithm that fused the state vector of linear system with arbitrary correlated noises is developed. The sampling period of the fusion center is defined as the lease common multiple of multi-sensors measurement period, so the same number of the local state estimation is transmitted to the fusion center every sampling period and the computation burden of the tracking filter is alleviated. The correlations between measurement noises and processing noises are added to the gain matrix of tracking filter, so information that can describe multi-sensors fusion systems is increased. Simulation results show that the proposed algorithm is better than others in which the measurement noises and processing noises are assumed to be uncorrelated, and the accuracy of trajectory track is improved.