基于海量物联网数据并行压缩感知研究.pdfVIP

* 基于海量物联网数据的并行压缩感知研究 张永平，张功萱，朱昭萌，张巍 （南京理工大学计算机学院 南京 210094 ） 摘 要：物联网是实现人与物、物与物互连的网络。在大规模的物联网中，大量智能感知系 统将产生海量的数据，这些数据的采集、处理与存储将对智能感知系统的寿命和物联网的性 能产生很大影响。压缩感知是一种能够在采样的同时实现数据压缩的采样方法，它通过降低 采样率减少采集的数据量，从而减轻网络的通信负担，但是压缩感知算法的计算复杂度很高、 对信号的适应性较差。本文提出使用压缩感知理论处理海量物联网数据，并尝试采用并行处 理的方法提高压缩感知的计算速度；同时针对压缩感知对信号适应性差的问题，给出了使用 冗余字典构造变换基的方法。 关键词：物联网；海量数据；压缩感知；优化重构 中图法分类号：TP Research of Parallel Compressed Sensing for Mass Data in Internet of Things Yongping Zhang, Gongxuan Zhang, Zhaomeng Zhu, Wei Zhang （School of Computer Science and Technology, Nanjing University of Science and Technology, Nanjing, China, 210094 ） Abstract: Internet of Things is the network of interconnection between people and things, and between things themselves. Mass data are usually collected, processed and stored in Internet of Things and this will adversely affect Internet of Things. Compressed sensing is a new sampling mean that data sampling and compressing can be done simultaneously. Compressed sensing can significantly lower the network communication burden by reducing sampling rates of sensors, but it is nonadaptive and its algorithm has high computational complexity. In this paper, we use compressed sensing theory to process mass data in Internet of Things and put forward the idea that is parallel processing of compressed sensing algorithm. At the same time, we introduce redundant dictionary into compressed sensing for increasing the flexibility. Keywords: Internet of Things; compressed sensing; parallel processing; optimal recovery 1. 引言 物联网（Internet of Things，IoT ）是当前最热门的科技词汇之一，其概念是麻省理工学 院的Kevin Ashton 于1999 年提出来的[1] 。在物联网中，物体的信息由智能感知系统（如： *本课题得到“江苏省973 计划”（No. BK2011022 ）、国家自然科学重点基金（No.）以及南 京理工大学重点基金（No. 2011YBXM18 ）资助。张永平，1979 年生，男，博士在读，