论文部分内容阅读
大规模的地理环境监测,以及用来传输与处理数据的物理基础设施无法和监测区域的规模保持同样的增速,使得不可靠链路下的数据采集与处理呈现出一种饱和流状态,无线传感器网络的能力看似难以稳定.另一方面,尽管理想网络模型的计算结果足够精确,然而,由于与实际应用偏差甚大,使得网络用户无法充分地分析和利用从工业现场所获得到的网络感知数据,并且没有针对网络规模和性能对数据聚融的影响进行分析.为此,本文提出以“过渡区”作为工业现场仿真的假设条件,并在此基础上提出了一种面向实际应用的数据聚融算法,即基于可信度的数据聚融算法(R算法).在具体设计R算法过程中,选用聚集和操作符SUM为例,通过对网络提供出的近似聚集和加以自动分析、综合,针对相对误差界限ε,计算出近似聚集和的可信度的下限η;并将近似聚集和、参数η一同提供给用户,在为用户提供网络概要信息的同时,还提供了参数η作为对信息可信度的判断,以便指导用户对数据聚集结果进行深度处理和提高网络的感知性能.仿真实验描述了过渡区内由于信噪比导致的链路不可靠所引起的η的变化规律;讨论了网络性能和规模对η的影响,随着网络运行周期的增加和网络规模的增大,η的值将逐渐靠近0;从而为WSNs从理论模型投入实际工业应用提供了理论依据和经验公式.
Large-scale geo-environment monitoring, as well as the physical infrastructure used to transmit and process data, can not maintain the same growth rate as the size of the monitored area, allowing data acquisition and processing over unreliable links to exhibit a saturated flow state. Wireless On the other hand, although the calculation results of the ideal network model are accurate enough, the deviation from the practical application makes the network users unable to fully analyze and utilize the network perception obtained from industrial sites Data, and did not analyze the impact of network size and performance on data aggregation.Therefore, this paper proposes to use “transition zone ” as the assumption of industrial field simulation, and on the basis of this proposed a practical application (R algorithm) based on the credibility of the data fusion algorithm.In the specific design of the R algorithm, the use of aggregation and operator SUM as an example, through the network to provide approximate aggregation and to be automatically analyzed , The lower limit of the credibility of the approximate aggregate sum is calculated for the relative error limit ε, and the approximate aggregate sum is provided together with the parameter η While providing network summary information to users, it also provides parameter η as a judgment of the credibility of information, so as to instruct users to process the data aggregation results in depth and improve the perceived performance of the network.A simulation experiment is conducted to describe the internal The change rule of η due to the unreliable link caused by signal-to-noise ratio; the influence of network performance and scale on η is discussed; with the increase of network operation period and the increase of network size, the value of η gradually approaches 0 This provides a theoretical basis and empirical formula for WSNs to invest in practical industrial applications from theoretical models.