摘要
现代城市中,传感器设备无时无刻地收集和释放着各式各样的城市数据,而这些数据可能会因环境干扰、设备故障或人为篡改而变得异常。针对城市数据会面临数据异常的问题,定义了一套适用于异构数据的提取协议,并基于提取协议设计了一种基于高斯隶属度的分布式异常数据分级过滤算法。最后通过仿真实验验证了该算法能够准确过滤异常数据,从而提高城市计算效率,降低网络开销。
In modern cities,sensor devices collect and release all kinds of urban data all the time.These data may become abnormal due to environmental interference,equipment failure or human tampering.Aiming at the problem of anomaly data filtering,a set of extraction protocols suitable for heterogeneous data was defined.Based on the extraction protocol,a distributed anomaly data classification filtering algorithm was designed based on Gaussian membership.Simulation experiments verified that the proposed algorithm could accurately filter abnormal data,so as to achieve the purpose of improving urban computing efficiency and reducing network overhead.
作者
王誓伟
徐晓斌
梁中军
WANG Shiwei;XU Xiaobin;LIANG Zhongjun(Center of Future Network Innovation, Beijing University of Technology, Beijing 100124, China;National Meteorological Information Center, Beijing 100081, China)
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2021年第9期2525-2531,共7页
Computer Integrated Manufacturing Systems
关键词
城市计算
物联网
大数据
异常过滤
移动边缘计算
urban computing
Internet of things
big data
abnomaly filtering
mobile edge computing