中文摘要：

移动Agent（mobile agents,简称MA）技术在无线传感器网络的异常事件监测中广泛应用.基于异常事件的性质,数据中心（即sink 点）派出实现特定功能的MA 到异常目标位置附近收集传感器节点（数据源节点）的感应监测数据.与传统的客户端-服务器传输模型不同,MA技术往往能够对收集到的感应数据进行压缩和融合,从而大大减少了在网络中传播的数据流量,降低了数据传输的能量消耗.通常情况下,MA 经过的源节点越多,其收集的数据量越大.通过对目前已有的MA 行程规划算法进行研究,提出了实现Agent 数据分流的传输技术.该技术基于分流规则,决定是否提前把MA 已收集到的数据传回sink,而不携带数据的MA 继续访问余下的传感器节点.该Agent 数据分流技术解决了由于MA 携带大量数据访问源节点而造成的源节点能量消耗过快的问题,适用性强.大多数目前常用的MA 行程规划算法都可以采用该技术来提高性能.Agent 数据分流的最终目标是减少源节点传输的数据量,尽可能地延长源节点的工作寿命,从而更长时间地实现对异常目标的监控.

英文摘要：

Mobile agents （MA） have been widely applied in detecting abnormal events in wireless sensor networks. Based on the characteristics of the event, the data center （i.e., the sink） sends a specific MA to collect the data sensed by the sensors （acted as source sensors） in the nearby area of the target event. Different from the traditional client-server transmission model, the MA technology can compress and aggregate the sensed data and reduce the amount of data and the energy consumption of sensors for data transmission. When MA passes more source sensors, the size of the collected data becomes larger. This paper investigates the existing MA itinerary planning algorithms, and proposes an agent data separation strategy. The strategy can determine whether to send the collected data to the sink in advance based on a data separation rule. If data separation is performed, the data collected by the MA will be sent to the sink, while the MA only with the head will continue visiting the rest source sensors. Because the data carried by the MA is reduced, the proposed agent data separation strategy conserves the energy consumption of sources sensors. The strategy is a general scheme and can be applied in most of the existing MA itinerary planning algorithms to enhance their performance. The ultimate goal of the proposed strategy is to reduce the data transmitted by source sensors and extend their working lifetime for monitoring the targets.