节点文献
多点监测自适应网内数据融合技术的研究及应用
【作者】 温冬伟；
【导师】 王平立；
【作者基本信息】 南京理工大学 ， 计算机系统结构， 2012， 硕士
【摘要】 多点监测无线传感网络中,由于节点能量有限和数目众多的特点导致网络的寿命受到限制。网络中数据包的无线收发消耗了节点的大部分能量,在无线传感网络内部的传感数据汇集过程中进行数据融合(网内数据融合)可以降低网络中数据传输量、降低网络整体功耗和提高网络寿命,但它又会带来额外的数据包延迟和监测精确度降低的副作用。本文详述了无线传感网络体系结构及其协议标准,讨论了网内数据融合技术及其在多点监测系统中的应用,对数据融合技术、聚类分析思想和模糊集合理论进行了分析,针对经典模糊聚类算法步骤,给出一种改进的聚类中心初始化方法,通过算法仿真和对比分析证明其改进效果,并将改进的模糊聚类算法应用于数据融合中,设计了分层的网内数据融合结构和自适应参数调整的融合模型。融合模型应用在无线多点监测系统中,随监测目标的变化自适应地调整网内融合的各参数,从而使监测的精确度自适应于监测目标的状态,系统在监测目标时降低数据的传输量、在监测目标异常时及时调整提高监测精确度,在大幅度降低网络整体功耗的同时又最大程度地减少因融合而产生的监测精度损失,使得监测系统能够及时准确的捕获每一次异常的发生,尤其适用于预警监测应用。论文最后通过分析煤矿自燃过程、自燃释放气体的浓度推测自燃状态的方法,针对煤矿自燃监测系统详细描述了模型的应用原型,设计了系统结构、各节点软件工作流程。
【Abstract】 Due to the large number of nodes and limited energy, the multi-point monitoring wireless sensor networks usually have a limited network lifetime. Wireless network packets sending and receiving consume most of the node’s energy. If aggregate data within the the wireless sensor network data collection(In-network Data Aggregation), it will reduce the amount of data transferred in the network, reducing the network’s overall power consumption and increase network lifetime. But it will bring additional packet delay and monitoring accuracy reduces.This paper detailed the wireless sensor network’s architecture and protocol standards, discussed the in-network data aggregation and its techniques in the application of multi-point monitoring system, analyzed cluster analysis algorithms and fuzzy set theory ideas, discussed classical steps of the fuzzy c means clustering algorithm, given an improved method of cluster centers initialization, verify the effectiveness of the improved algorithms through comparative analysis, and applied the improved fuzzy c means clustering algorithm to data aggregation, given a design of the hierarchical in-network data aggregation structure and an adaptive parameter adjustment model.Applied this system model to wireless multi-point monitoring system, it can adaptively adjust the aggregation parameters to the changing of monitoring goal in the network. It reduce the amount of data transfer when the system monitor the target, and increase monitoring precision degree when the monitoring target abnormal. It can reduce the overall network power consumption significantly and minimize the precision loss due to the data aggregation. And it can making the monitoring system to capture every unusual occurrence timely and accurate. At last, the paper analyzed the process of coal spontaneous combustion and the gases of coal spontaneous combustion, detailed the method of use gases to speculate the coal spontaneous combustion state, discribled the application of the prototype model, given a design of the system architecture and each node’s software workflow.