WSN中一种基于数据融合的Mobile Agent路径规划方法

摘要

MA（Mobile Agent）技术作为一种分布式中间件的方法，它很适合应用于WSN中自主性的数据融合和能量均衡方面。但是现有的一些MA路径规划方法确定的路径只考虑节点间的物理距离，此时网络会出现数据负载不均衡、延迟和安全性等一系列问题。针对上述问题，提出一种基于MA协议的路径规划方法（LDLM），该算法采用K-means聚类算法对网络分簇，然后由Sink节点根据每个簇节点规模派发若干个MA；并且在确定每个MA节点访问组时，综合考虑了节点采集数据量的规模和MA可用存储空间。最后，根据每个MA节点访问组，使用LCF算法确定MA对其节点访问组的访问路径。仿真实验结果表明：该算法在能量消耗、网络生命周期、任务工作周期三个方面性能表现优于现有的一些算法。%As a distributed middleware,Mobile Agent technology is suitable for autonomic data fusion and energy balance in wireless sensor networks. Several heuristic methods have been widely applied to the way the Itinerary planning of Mobile Agent,but some of these methods just consider the geographical distance among nodes as the unique factor when planning the Itinerary. There will cause the emergence of the data load unbalancing,large delay and security problems when WSN uses these methods. For tackling above problems,the Larger Data size in the Larger Memory(LDLM)algorithm is proposed. In this proposed scheme,we don’t only consider the geographical distance among nodes but take into account the data size from source nodes. Extensive simulation experiments show that the LDLM algorithm behaves better performance than other approaches on these three aspects consumption en⁃ergy,life cycle and task duration.