一种基于分步式滤波的异步航迹融合算法

An Algorithm of Asynchronous Track Fusion Based on Step Filtering

针对分布式多传感器数据融合系统,改进了一种多传感器异步航迹融合算法.该融合算法在扩展记忆因子递推最小平方(EFRLS)估计形成本地航迹的基础上,首先利用最小二乘法将局部航迹统一到融合中心的融合时间点,然后使用分步式滤波融合算法进行状态估计.仿真实验结果表明,这种融合算法在噪声方差未知的情况下是有效的,其精度接近集中式融合算法.

Concerning for distributed multi-sensor data fusion system, a multisensor asynchronous track fusion algorithm was improved. In this algorithm, first, the local tracks were synchronized to the fusion point of the fusion center by using the least mean square （LMS） algorithm. Next, an algorithm of asynchronous track fusion based on step filtering was employed to implement the estimation of states. Simulation results show that this fusion algorithm is effective when the noise variance is unknown, and its accuracy approaches to that of the centralized processing algorithm.