具有一步随机滞后和丢包多传感器系统分布式递推融合预报器

Distributed recursive fusion predictor for multi-sensor systems with one-step random delays and packet dropouts

研究了具有一步随机滞后和丢包多传感器系统的分布式递推融合估计问题。利用满足伯努利分布的随机变量描述传感器到估计器的随机滞后和丢包现象。通过定义新变量,将原系统等价地转化为随机参数化系统。基于局部最优线性估值,局部估值之间的互协方差阵,以及先验融合估值和局部估值之间的互协方差阵,提出了分布式递推融合预报算法。给出稳态预报器存在的一个充分条件。通过仿真验证其有效性。

Distributed recursive fusion estimation problem for multi-sensor systems with one-step random delays and packet dropouts is concerned.Bernoulli distributed random variables are used to model the phenomena of random delays and packet dropouts from sensors to estimators.By defining some new variables,the original system is equivalently transformed into a stochastic parameterized system.Based on local optimal linear estimates,cross-covariance matrices between local estimates,and between prior fusion estimate and local estimates,a distributed recursive fusion prediction algorithm is presented.A sufficient condition for the existence of the steady-state predictor is given.An example verifies the effectiveness of the algorithms.