中文摘要：

针对无线传感器网络的最小二乘定位算法抗差性的不足，提出了一种基于时空滤波（STY）的抗差性加权最小二乘（WLS）节点定位算法——吣。该算法基于空间域滤波的数据一致性检测算法利用相邻节点间必须满足的几何约束关系，采用优化矩阵操作，剔除粗差邻居节点，其计算复杂度为多项式的平方。通过使用具有2步收敛特性的DFP算法，最小化目标代价函数，实现节点的快速定位。实验结果表明，在均匀网格拓扑或各向异性C型网格拓扑下，该算法均可有效识别和剔除测距低估粗差点，其定位精度明显优于未进行空间一致性检测的加权最小二乘定位算法，当网络平均连通度较低时，该优势表现得尤为明显。

英文摘要：

This paper presents STLS, a robust weighted least squares localization algorithm based on spatio-temporal filter for wireless sensor networks to improve the positioning resilience of the least squares scaling when there are outliers in the ranging measurements. Its ranging consistency check scheme is based on geometric ranging inequations to which neighboring nodes must obey. By building matrixes, the STLS speeds up identifying and eliminating the outliers of the ranging measurements only with the computing complexity of quadratic polynomial, and employs a 2-step-convergence DFP algo- rithm to minimize the cost function to fast localize the unknown nodes. For received signal strength based range measurements, regardless of uniform grid topology or irregular C-shaped grid topology, the extensive simulation results confirm that STLS can detect and delete the underestimated outliers effectively and outperforms the traditional weighted least squares scaling which has no consistency check and demonstrates more advantages when there are sparse range measurements.