基于传感器丢包率不确定性预测的分布式H∞滤波算法

Distributed H∞ filtering algorithm based on uncertainty prediction of packet loss rate of sensor

针对无线传感器网络(WSNs)在测量数据丢包不确定性情况下的分布式H∞滤波问题,提出一种基于传感器丢包率不确定性预测的分布式H∞滤波算法。通过设计一种线性全阶滤波器,使得滤波误差能够收敛在均方渐近为零,同时抑制外界的干扰造成的H∞滤波衰减。根据滤波参数符合伯努利分布随机分布的原理,利用Lyapunov函数的方法,通过寻找最佳的滤波参数来保证在丢包率不确定的情况下对真值预测的随机稳定性。实验仿真结果表明:当观测数据存在不确定性丢包时,该滤波算法能发挥有效的滤波效果。

Aiming at distributed H∞ filtering problem of measurement data loss in case of uncertainty in wireless sensor networks（ WSNs ）,a distributed H-infinity filtering algorithm based on sensor uncertainty prediction of packet loss rate is proposed. By designing a linear full-order filter,so that filtering error can converge in mean square asymptotically to zero,while suppressing H∞ filtering attenuation caused by external interference. According to Bernoulli distribution filtering parameters consistent with the principles of random distribution,by searching the optimal filtering parameters to ensure random stability in the case of packet loss rate uncertainty of true value prediction,using the Lyapunov function. Simulation results show that this filtering algorithm can play an effective filtering effect if there is uncertainty observational data loss.