无线传感器网络中基于潜在博弈的分布式节点定位  

Distributed Node Localization Based on Potential Game in Wireless Sensor Networks

作　　者：贾杰[1,2] 张桂园 陈剑[1] 王兴伟[1] 赵林亮[1] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110819 [2]沈阳理工大学,辽宁沈阳110159

出　　处：《电子学报》2014年第9期1724-1730,共7页Acta Electronica Sinica

基　　金：国家自然科学基金(No.61173153;No.60903159;No.61070162;No.71071028);国家杰出青年科学基金(No.61225012);中央高校基本科研业务费专项资金资助(No.N110404014;No.N110318001;No.N110204003;No.N120104001);中国博士后科学基金(No.20110491508;No.2012T50248);高等学校博士学科点专项科研基金优先发展领域资助课题(No.20120042130003);沈阳理工大学计算机应用技术重点学科开放基金资助课题(No.4771004kfx06)

摘　　要：分布式定位是无线传感器网络研究中的热点问题.传统的分布式定位机制存在定位精度低,算法求解复杂等问题.为此,以与邻居节点集距离误差和作为效益函数,提出基于博弈论的分布式定位模型.给出了该博弈模型为潜在博弈的形式化证明,并从理论上证明了纳什均衡的存在性及最终收敛解的有效性.进一步,仅通过与邻居节点进行策略信息交互,提出基于潜在博弈的分布式定位算法.最后设计了博弈策略空间的决策机制与未知节点的升级机制,以避免陷入局部最优并加速收敛至全网最优解.一系列仿真实验验证了所提算法的有效性.Distributed node localization is an important issue in wireless sensor networks .However ,traditional distributed lo-calization algorithms have the drawback of low localization accuracy and high processing complexity .In response to these problems , a distributed localization model based on game theory is presented ,where the utility function for each participant is defined as the sum of neighbor distance error .Formally ,the proposed localization game is proved as a type of potential game .Through theoretical analysis ,the existence of Nash Equilibrium and the validity of the final converged solution are testified .Furthermore ,a distributed localization algorithm based on game theory is proposed ,where each sensor exchanging information only with its neighbors .Finally , to avoid converging in local optimum and accelerate the convergence speed ,the strategy space determination and unknown nodes el-evation are developed .Extensive simulation results are performed to demonstrate the effectiveness of our proposed algorithm .

关 键 词：无线传感器网络 分布式定位 潜在博弈 效益函数 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]