Abstract: Node localization in wireless sensor networks is a basic but very important research field. In practical application, sensor nodes are distributed randomly resulting in node density disproportion in the region. The existing localization algorithms of node are not sensitive to distribution density. Hence, if the algorithm applies the same strategy to locate nodes in regions with different node density, it will cause position accuracy low in dense area and location rate decreasing in sparse regions and the beacon node energy is not to maximize utilization. This paper puts forward a spanning tree algorithm of beacon based on node distribution density called GBT. Nodes in beacon group are traversed by planned path to complete node location. The simulation results show that the improved algorithm has better performance on decreasing time expend for locating node, enhancing accuracy of node location by taking advantage of beacon efficiently compared with previous algorithms.

Key words: wireless sensor networks, sensor localization, dynamic beacon node, route planning, node density, Generate Beacon-based Tree（GBT） algorithm

摘要： 节点定位是无线传感器网络中一个基础但十分重要的研究方向。实际应用场景中，传感器节点大多被随机部署，分布往往疏密不均。现存的定位算法对节点的分布密度没有敏感性，如果算法在节点密集区域和稀疏区域使用相同的定位策略，就会造成密度大的区域定位精度低，分布相对稀疏的区域定位率低，信标节点的能量得不到最大化利用等问题。针对这些问题，提出了一种基于节点密度进行定位的生成信标树算法（GBT）。信标节点组沿着规划好的路径对节点进行遍历，实现节点的全定位。通过与其他规划动态信标节点路径算法比较，证明了GBT算法在定位时间、定位精度和对信标节点能量的充分利用上均有所改善。

关键词: 无线传感器网络, 传感器节点定位, 动态信标节点, 路径规划, 节点密度, 生成信标树算法