基于TDOA测距模型的SOCP和Taylor混合定位方案

TDOA ranging based Second Order Cone Programming and Taylor Hybrid localization

针对无线传感网络WSNs(Wireless Sensor Networks)的定位问题,提出基于TDOA(Time Different of Arrival)测距模型的SOCP(Second order cone programming)和Taylor混合定位方案,记为SOCP+Taylor。SOCP+Taylor方案首先分析了网络可定位性,并结合刚论(rigid),提出了评判节点可定位的条件。然后,建立TDOA测距模型,并用最大似然估计建立距离测量值的最大似然函数,再通过松弛约束理论将非凸优问题转换成凸优问题,引入惩罚因子,进而利用SOCP估计节点位置,并将此节点位置作为泰勒级数展开法Taylor迭代的初始值,最后,利用Taylor估计节点的最终位置。在不同参考节点数目以及变化的噪声环境下对算法进行仿真。仿真结果表明,提出的定位方案具有高的定位精度,定位误差逼近于CRLB,同时分析了惩罚因子对定位精度的影响,并确定了惩罚因子的最佳取值区域。

For localization in Wireless Sensor Networks（ WSNs）,the TDOA ranging based Second Order Cone Programming and Taylor Hybrid localization scheme is proposed in this paper,which is marked as SOCP ＋ Taylor scheme. Firstly,in view of network localizability,making use of rigid theory,the new localizability criterion is proposed in this paper. After that,establish TDOA-based ranging model and function of measurement by using maximum likelihood estimate criterion. Non-convex optimal problem is converted into convex optimal problem by relax. Using a penalty term,the initial position of node is estimated by the SOCP,which is considered as the initial guess for Taylor method. Finally,the initial position of node is estimated by the Taylor. Under various number of reference node measurement,the simulations on localization performance is done. Simulation results show that the performance of proposed localization scheme attain CRLB,and determine the best value range of penalty factor.