Abstract Wireless Sensor Networks grow rapidly due to the sensing and scanning capabilities of nodes. The sensors are capable to move randomly and send periodic reports, hence they must be processed efficiently. This requirement increases the indexing burden, in particular, indexing the mobile sensors’ data. The grid-based indexing techniques in random mobile sensors suffer from high dependability on packet attributes and the necessity of registering the current cluster location of each sensor and assign its destination periodically. Therefore, this work proposes a framework called Dynamic-Coalition to efficiently improve the indexing packets of random mobile sensors. The proposed framework alleviates the disseminated packets by constructing dynamic-coalitions using relevancy-measure. The dynamic-coalition framework is superior in terms of average routing overhead and packet delivery ratio. It can significantly outperform R-tree, Decomposition-tree, and Grid-based DBSCAN with Cluster Forest in terms of index building time overhead, the number of traversed nodes, and space cost overhead.

中文翻译：

---

一种用于无线传感器网络中数据传播的高效索引框架

摘要 无线传感器网络由于节点的感知和扫描能力而迅速发展。传感器能够随机移动并定期发送报告，因此必须对其进行有效处理。这一要求增加了索引负担，特别是索引移动传感器的数据。随机移动传感器中基于网格的索引技术存在对数据包属性的高度可靠性以及需要注册每个传感器的当前集群位置并定期分配其目的地的问题。因此，这项工作提出了一个称为 Dynamic-Coalition 的框架来有效地改进随机移动传感器的索引数据包。所提出的框架通过使用相关性度量构建动态联盟来减轻传播的数据包。动态联盟框架在平均路由开销和数据包传输率方面更胜一筹。在索引构建时间开销、遍历节点数和空间开销开销方面，它可以显着优于 R-tree、Decomposition-tree 和基于 Grid 的 DBSCAN with Cluster Forest。