Information filtering method adopting parallel fusion mechanism

The invention discloses an information filtering method adopting a parallel fusion mechanism. According to the information filtering method, a local posteriori estimation mean value and a covariance which are updated in real time are applied to a local unscented information filter, so that the unscented information filter and a weighted average consistency filter run in parallel, normal operationof the algorithm in a random dynamic communication topology is ensured, an average consistency filter is improved, and a secondary small characteristic root of a communication topology Laplace matrixis maximized, so that the algorithm has better stability, accuracy rate and convergence rate, and has higher engineering application value. 本发明公开了一种采用并行融合机制的信息滤波方法，通过将实时更新的局部后验估计均值和协方差用于局部无迹信息滤波器，使无迹信息滤波器和加权平均一致性滤波器并行运行，保证算法在随机动态通信拓扑正常运行，对平均一致性滤波器的改进和通信拓扑拉普拉斯矩阵次小特征根的最大化，使算法具有更好的稳定性，准确率和收敛速率，具有较高的工程应用价值。