【作者】 李奕；

【导师】 戴居丰；

【作者基本信息】 天津大学 ， 通信与信息系统， 2012， 博士

【摘要】 物联网是由多项信息技术融合而成的新型技术体系,被称为继计算机、互联网之后,世界信息产业的第三次浪潮。作为连接所有“物-物”关系的泛在网络,物联网还处于“概念”阶段,虽然已有部分成熟技术可供应用,但是,更多相关领域的研究还没有展开。因此,信道模型、协作传输方法以及信息压缩处理都是其研究的重要内容,与物联网能否进入实用阶段有着密切的联系。本文主要工作与贡献包括以下几个方面:1.提出了基于物联网市区环境的改进Clarke模型。依据物联网市区环境的特点,克服了Clarke理论模型虽能精确描述信道但计算量过大,Jakes模型及很多改进模型精简了计算量但统计特性不好的弊端,提出一种改进Clarke模型。通过统计特性的计算,表明改进模型在比理论模型减少约一半计算量的基础上,保持了广义平稳态,且在有限N值的情况下,可以得到更好的高阶统计特性。通过仿真结果的比较,改进模型的一、二阶统计特性均比Wu模型(一种改进Jakes模型)更贴近理论模型,具有更好的独立性。2.提出了基于物联网远郊环境的改进Rician模型。通过推导分析和仿真证明了该模型符合Rician分布特点,是广义平稳的并具各态历经性,比较了Rician因子取不同值时统计特性的变化。改进Rician模型的特点是它的互相关函数尤其是相邻信道的互相关函数非常小,具有非常好的独立性。3.研究了协作分集技术,计算了直接传送方式以及协作分集技术(含放大转发和解码转发方式两种方式)的信噪比及误码率,分别以Matlab内建仿真信道模型、1中提出的改进Clarke模型和2中提出的改进Rician模型为底层通信信道,搭建了协作分集系统仿真平台,进行了误码率的验证。仿真结果表明了改进模型的可信性和有效性,并得出协作分集技术更适用于物联网环境的结论。4.提出了两种针对物联网物理标识语言(PML)文件的无损压缩算法-改进Ⅰ型和Ⅱ型Huffman编码算法。Ⅰ型算法为基于Huffman编码的整元素压缩法,属于最优化编码;Ⅱ型算法针对Ⅰ型算法中可能出现的内存空间不足情况,以平均码字长度的增加换取占用内存空间的减少。通过比较标准Huffman、改进Ⅰ型和改进Ⅱ型算法在处理不同大小文件时的压缩情况,证明两种改进算法的压缩比都大于标准Huffman编码,能为物联网提供更为高效的传输效率。而且,改进算法弥补了Huffman编码没有错误保护功能的缺陷,在不增加传送数据总量的同时,为解压译码提供了一定的检错能力。更多还原

【Abstract】 Internet of Things (IOT) is a new technology system which is combined by a lot of information technologies. It’s known as the third revolution of IT industry after the computer and internet. As the ubiquitous network which contains kinds of object-object connections, IOT is still in the concept stage. There’re already some mature technologies available for reference. However, little research work in this area has been started. Therefore, the research of channel model, cooperative diversity and data compression are all of great importance, and are critical for realizing the practical stage of IOT. The main contributions of this thesis are as below:1. An improved Clarke model is proposed for the urban environment in the IOT. The theoretical Clarke model can accurately describe the channel, but need large quantity of calculation. And the Jakes model has less calculation but with bad statistical properties. These shortcomings are improved in the new model. The caculation of statistical properties indicates that the improved model is wide-sense stationary while reducing nearly half of the computation. And with a finite value of N (number of sinusoids), it has a better high-order statistics. The simulation result shows that the new model’s first-order and second-order statistics are closer to the theoretical values than Wu model (an imprved Jakes model). Its performance of independence is good.2. An improved Rician model is proposed for the suburban enviroment in the IOT. The derivation and simlution results prove that this new model matches the Rician distribution, and is wide-sense stationary. It is also simulated with different values of Rician factor for checking the statistical properties. Its cross-correlation value is very small, especially for the adjacent channels, which insures a good performance of independence.3. We study the technique of cooperative diversity, and caculate the SNR and BER of direct transmission, AF (Amplify and Forward) and DF (Decode and Forward) with cooperation. The Matlab simulation platforms for the embedded channel, the improved Rayleigh channel in 1 and the improved Rician channel in 2 are all created to validate their BER performance. Simulation results prove the credibility and effectiveness of this improved model, which indicates that cooperative diversity is more suitable for the environment of IOT.4. Two types of improved Huffman compression algorithms (ⅠandⅡ) for PML file in the IOT are proposed. TypeⅠcompresses the whole element part based on the Huffman coding and is most optimized. TypeⅡis designed for the insufficient memory situation, by seperating the weight assignment for element and data parts, and its memory requirement is lessen with increasement of average code word length. With comparison of standard Huffman coding, the improved typeⅠand typeⅡalgorithm’performance of handling different size files, we prove the compression ratios of these two modified algorithms are much bigger than the standard Huffman coding, and can provide transmission efficiency for IOT. Besides these, the improved algorithms make up the Huffman coding’s defect of no error checking. Without increasing the amount of data transmission, new algorithms provide error detection function for the decompression and decoding.更多还原