不确定观测线性离散随机系统白噪声估值器

White Noise Estimators for Linear Discrete-Time Stochastic Systems with Uncertain Observations

在网络控制系统和传感器网络中,可能的传感器观测数据丢失使得系统的观测具有不确定性。应用新息分析方法,对传感器具有数据丢失的不确定观测线性离散随机系统,提出了统一和通用的白噪声估计算法,包括输入白噪声估值器和观测白噪声估值器。可统一处理传感器具有数据丢失的白噪声的最优滤波、预报和平滑问题。同时,给出了稳态白噪声估值器和相应的Wiener白噪声估值器。当没有数据丢失时,所得的结果恰是以往基于完整观测数据的白噪声估值器。仿真研究验证了算法的有效性。

In networked control systems and sensor networks,possible observation data losses of sensors make the observations of systems uncertain.The unified and universal white noise estimation algorithms are developed for linear discrete-time stochastic systems with uncertain observations due to data loss of sensors by applying the innovation analysis approach,which include input white noise estimators and observation white noise estimators.They can handle the optimal filtering,prediction and smoothing problems for white noise in a unified framework in the case of observation data loss of a sensor.The steady-state white noise estimators and corresponding Wiener estimators are also given.When there is no data loss,the proposed results are just reduced to the previous white noise estimators based on complete observation data.A simulation verifies the effectiveness of the proposed algorithms.