一种基于数据质量的异构多源多模态感知数据获取方法

A Heterogeneous Multi-Source Multi-Mode Sensory Data Acquisition Method Based on Data Quality

近年来,物联网由于其广阔的应用前景得到快速发展,感知设备的种类也越来越丰富.并且很多应用中均通过部署多个相互独立的异构数据源对监测对象的多维属性进行度量,由此得到异构多源多模态感知数据.但由于硬件设备固有的限制以及环境噪声等因素的影响,感知设备不可避免地存在漏读、多读、错读等现象,导致感知数据的数据质量降低.而数据是信息的载体,其能否准确反映物理世界的真实信息是决定其在上层应用中发挥效用的关键.基于此,该文提出一种基于数据质量的异构多源多模态感知数据获取方法.首先定义了数据质量的3个方面:准确性、完整性和一致性;进而对数据质量在这3个方面进行建模,提出评价异构多源多模态感知数据综合数据质量的方法;最后提出基于数据质量的感知数据获取算法,根据用户给定的任意精度,选择部分数据源用于数据传输,在保证数据质量的前提下尽可能地减少网络资源的消耗,并通过大量实验验证了算法的高效性和可用性.

In recent years, with the rapid development of the IOT （the Internet of Things）, the type of sensing devices has become increasingly diversified. And in many applications, the sensor network usually consists of a host of mutually independent data sources, which can be used to monitor measured objects from multiple dimensions thereby obtaining the heterogeneous multisource multi-modal sensory data. However, there still exist false negative readings, false positive readings and wrong readings which reduce the data quality due to the inherent hardware limita tions and environmental interference, etc. Therefore, we propose a heterogeneous multi-source multi-mode sensory data acquisition method based on Data Quality （DQ）. We first define the data quality in terms of three aspects--accuracy, integrity, and consistency. Then, by modeling these aspects respectively, we propose metrics to estimate the comprehensive data quality method of heterogeneous multi-source multi-mode sensory data. Finally, a data acquisition method is presented based on data quality, which selects a part of data sources for data transmission according to the given precision. This method aims at reducing the consumption of the sensory network on the premise of the data quality guarantee. An extensive experimental evaluation demonstrates the efficiency and effectiveness of the algorithm.