具有随机多丢包的ARMA信号最优线性估计

Optimal linear estimation for ARMA signals with stochastic multiple packet dropouts

研究网络环境下具有随机丢包的自回归滑动平均(ARMA)信号的估计问题,其中丢包现象通过一个满足Bernoulli分布的随机变量描述.通过ARMA模型与状态空间模型的转化,将具有丢包的ARMA信号估计问题转化为具有丢包的状态空间模型的状态和白噪声估计问题.利用射影理论分别给出线性最小方差最优线性状态估值器和白噪声估值器,进而获得ARMA信号估值器.仿真结果表明,当存在数据丢失时,所提出的算法与以往基于完整数据的最优估计算法相比具有最优性和有效性.

The estimation problem for autoregressive moving average（ARMA） signals with stochastic packet dropouts in the networked environment is studied,where the phenomenon of the packet dropouts is described by a Bernoulli distributed random variable.The estimation problem for ARMA signals with packet dropouts is converted to the one for the state and white noise of state space model with packet dropouts by the transform from ARMA model to the state space model.Based on projection theory,the optimal linear estimators for the state and white noise are derived in the linear minimum variance sense respectively.Further,the estimators for ARMA signals are obtained.Simulation results show that,in the presence of packet dropouts,compared with the previous optimal estimation algorithms based on the complete data,the proposed algorithm has optimality and effectiveness.