中文摘要：

在森林防火、目标追踪、灾难预警、环境监测等应用中，需要通过定位算法对无线传感器节点进行三维定位。提出一种基于多维定标的无线传感器网络三维定位算法，结合RSS经验衰减模型和最短路径建立相异性矩阵，采用轻量级矩阵分解算法降低相异性矩阵分解的计算复杂性，并利用网络中存在的周期性消息将初始定位信息回送，在后台使用迭代优化算法对初始定位结果求精。仿真实验表明，在测距误差一定的情况下，该算法能够提高节点三维坐标的初始计算精度，经过集中式的优化求精后与MDS-MAP算法相比，能够明显地提高节点三维定位的精度。

英文摘要：

In wireless sensor networks, three-dimensional 3 D localization is very crucial for the applications of fire preventing, target tracking, disaster warning, and environmental monitoring, etc. This paper proposed a three-dimensional localization algorithm based on multidimensional scaling. Established the dissimilarity matrix by combining the experience attenuation model of RSS and the shortest path method. Reduced the computation complexity of decomposing dissimilarity matrix by a lightweight matrix decomposition algorithm. At the back end, refined the initial positioning coordinates by the iterative optimization algo- rithm. Simulation results show that the algorithm, in the case of stable measurement error, can refine the initial coordinates and achieve a higher precision of localization in three-dimensional space compared with the MDS-MAP algorithm.