多传感器异步采样系统的顺序融合

Sequential Fusion for Multi-sensor Asynchronous Sampling System

针对现有基于伪量测的异步融合算法存在实时性差、融合时刻计算负荷大以及人为引入噪声相关等问题,提出了一种新的基于采样点顺序离散化思想的顺序式异步融合算法.该算法首先将各个传感器的测量值在融合中心的坐标系中和时钟下进行映射统一;然后,选取融合周期内各采样时刻对连续状态系统进行顺序离散化,从而获得本周期内各采样点间的状态方程和相应的测量方程.最终,使用线性最小均方误差意义下最优的线性卡尔曼滤波器实现本周期内异步采样量测的顺序滤波融合.仿真分析表明,该算法和基于伪量测的异步融合算法相比具有较少的计算量、较好的实时性和较高的估计融合精度.

Aiming at the problems which exist in the asynchronous fusion algorithms based on pseudo-measurements, such as poor real-time performance, excessive computation burden at the fusion time, and man-made noises correlation etc., this paper proposes a novel sequential asynchronous fusion algorithm based on the idea of sequential discretization of the sampling points. Firstly, it maps and unifies all measurements in the reference frame and clock with fusion centre. Secondly, selecting every sampling time in the fusion period to discretize the continuous state system sequentially, we get the state equation and the relevant measurement equation between every sampling point in this period. Finally, using the best linear Kalman filter in the sense of linear minimun mean square error, the sequential filtering fusion of asynchronous sampling measurements in this period can be realized. Simulation analysis shows that the proposed algorithm has lower computational load, better real-time performance and higher accuracy compared with the asynchronous fusion algorithms based on pseudomeasurements.