论文部分内容阅读
针对稀疏无线传感器网络(WSN)中加权平均一致分布式无迹信息滤波(DUIF)算法估计次优和滤波效率较低的问题,提出一种考虑先验估计误差相关性的快速DUIF算法.采用加权统计线性回归(WSLR)方法线性化观测模型,以节点共享信息作为平均一致性算法输入,从而在极大后验估计中引入先验估计交互协方差信息;设计最优通信连接边权值并自适应修正状态加权矩阵,提高平均一致性算法收敛速率.仿真实验结果表明,所提出的算法能够有效应用于稀疏WSN目标跟踪.
To solve the problem of suboptimal and low filtering efficiency in weighted average uniform distributed traceless information filtering (DUIF) algorithm in sparse wireless sensor network (WSN), a fast DUIF algorithm considering the correlation of a priori estimation error is proposed. Statistical linear regression (WSLR) method is used to linearize the observational model, and the node shared information is input as the average consistency algorithm, so as to introduce a priori estimation of mutual covariance information into the maximum a posteriori estimation. Adapt to the modified state weighting matrix and improve the convergence rate of the average consistency algorithm.The simulation results show that the proposed algorithm can be effectively applied to sparse WSN target tracking.