未知时变区域内的移动传感器网络控制  

Mobile Sensor Network Control in Unknown and Time-varying Region

作　　者：闫茂德[1] 张建国 左磊 YAN Mao-de;ZHANG Jian-guo;ZUO Lei(School of Electronics and Control Engineering,Chang?an University,Xi'an 710064,China)

机构地区：[1]长安大学电子与控制工程学院,西安 710064

出　　处：《控制工程》2021年第2期299-305,共7页Control Engineering of China

摘　　要：针对时变区域内信息分布状态的估计问题,旨在通过动态控制传感器网络在目标区域内的采样位置,利用不同采样位置上的采样信息,实时动态地估计目标区域内的信息分布状态。具体地,首先设计出一种基于传感器采样变化幅度的动态采样机制,以减少测量信息中的冗余信息。然后将曲面拟合方法与卡尔曼滤波方法相结合,提出一种面向时变区域的动态信息分布估计算法。在此基础上,以信息分布的估计误差为目标函数,利用优化算法,设计移动传感器网络的最优采样位置控制律,进一步优化目标区域内信息分布状态的估计结果,并通过严格的理论分析证明所提估计算法的收敛性。最后,通过仿真实验的方法验证该控制系统与估计算法的有效性与可行性。In this paper,we consider the estimation problem for the information distribution in a time-varying region,which aiming at dynamically estimating the information distribution by adjusting the sampling positions of the mobile sensor network and using sampling information at different sampling positions.In details,we firstly present a dynamic sampling mechanism based on the sampling rate.It can reduce the redundant information in the estimation algorithm.Then,according to the Surface fitting techniques and the Kalman filter method,we propose a novel estimation algorithm for the time-varying information distribution.To further improve the estimation accuracy,we take the estimating errors as objective function and present a control scheme for the mobile sensor network by using the stochastic gradient descent.Furthermore,the convergence of proposed control system and estimation algorithm are strictly analysed.In final,numerical simulations are provided to illustrate the effectiveness and feasibility of proposed approaches.

关 键 词：时变区域 信息分布估计 移动传感器网络 动态采样机制 卡尔曼滤波 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置]