|Place of Conferral||中科院自动化研究所|
|Keyword||ACP方法 平行智能 物联网 低功耗广域网 正交频分复用|
物联网(Internet of things, IoT)已经成为继计算机、互联网之后，世界信息产业的第三次浪潮。随着物联网技术不断更新和发展，物联网逐渐呈现出异构化、碎片化和复杂化的趋势。同时，随着物联网应用场景越来越多样化，设计物联网解决方案时面临着难以建模，实践和经验不足等问题。平行智能有机结合人工系统(Artificial systems, A)、计算实验(Computational experiments, C)和平行执行(Parallel execution, P)，研究数据驱动的描述智能、实验驱动的预测智能和互动反馈的引导智能，是复杂系统的建模、分析和控制的有效途径。本文基于平行智能理论围绕物联网关键技术开展了一系列创新研究：面向智能交通和智能电网领域，以物联网真实应用场景为模型构建人工系统(A)，通过计算实验(C)分析物联网关键技术在多种应用场景中的性能；依据实验结果，本文提出多种新型物联网解决方案，进一步提升传统系统的性能，达到指导实际系统运行的目的(P)。本文的具体工作概括如下：
1. 面向城市交通传感网，提出一种新型的低功耗广域网（low power wide area network, LPWAN）参数优选方案。所提出的方案能够有效解决LPWAN应用于V2X（vehicle-to-everything, V2X）场景时所面临的时间选择性衰落问题。本文依据真实V2X环境的特点构建了虚拟的V2X人工系统，该系统包含车-车、车-路等六种典型场景。为进一步提取时间选择性衰落出现的规律，我们在人工系统中测试了多种LPWAN参数方案在各个V2X场景中的性能。经计算实验分析，本文设计了适用于各V2X场景的LPWAN参数优选方案。实验结果表明，提出的方案能够有效克服时间选择性衰落，可显著提升V2X环境中LPWAN系统的可靠性。
2. 面向复杂高速移动场景，提出了若干种车载多载波技术创新设计方案。复杂高速移动信道中存在严重的频率选择性衰落和多普勒效应，会全面影响多载波系统的性能，甚至诱发严重的载波间干扰（inter-carrier interference, ICI）。为对抗上述问题，本文基于平行智能理论，构建了面向车联网的人工系统，并对多载波技术进行一系列的创新设计：针对频率选择性衰落问题，本文提出一种基于交织的索引调制正交频分复用（index modulated orthogonal frequency division multiplexing, IM-OFDM）方案，该方案通过优化传统系统中的载波分组方式，可以获得频率分集增益，能够有效对抗频率选择性衰落；针对载波间干扰问题，本文提出了一种新型的载波间干扰自消除方案，通过有机融合IM-OFDM系统和ICI自消除思想，可以实现对抗载波间干扰的目的。计算实验表明，提出的两种方案更适用于复杂高速移动场景，相比传统方案可以获得更优的误码率性能.
3. 面向电力物联网领域，研究窄带物联网（narrow-band IoT, NB-IoT）应用于智能电网的技术优势和性能指标等问题。窄带物联网技术相比传统技术更适合智能电网的需求，具有覆盖广、功耗小、高可靠、支持海量连接的特点。本文将窄带物联网引入智能电网，不仅全面对比窄带物联网与传统电网传感技术的各项性能，而且从定量和定性两个角度针对窄带物联网适用的电网业务进行了综合分析、评估和分类。同时，基于平行智能理论，根据电网真实应用场景构建基于NB-IoT的人工系统，通过计算实验分析NB-IoT在城市、农村、山区等场景下的性能。实验证明NB-IoT在电网的各个典型通信场景下均获得出色的性能，能够良好地适配智能电网多种业务。
Internet of things (IoT) emerges as the third wave in the development of the global information industry after computer and Internet. With continuous upgrade of IoT technologies, IoT systems become more and more heterogeneous and complicated. Besides, with the increasing application diversity in IoT, it is difficult to model and simulate IoT systems due to lack of experience. Parallel intelligence and ACP approach can be seen as the effective ways to solve the aforementioned problems in IoT. ACP approach consists of artificial systems (A), computational experiments (C) and parallel execution (P), which aims to achieve the data-driven descriptive intelligence, the experiment-driven predictive intelligence and the interaction-driven prescriptive intelligence. Based on the ACP approach, parallel intelligence can transform a Cyber-physical-social system (CPSS) with uncertainty, diversity and complexity (UDC) into an intelligent system with agility, focus and convergence (AFC). It has been proved that ACP approach is an effective way to model, analyze, control and manage the complex systems. In this thesis, we perform a series of innovative researches on the key IoT technologies via ACP approach. First, the artificial systems are constructed based on the features of actual systems. Second, numerous calculation experiments are executed on the artificial systems to analyze the performance of IoT technologies in various applications. Two main application fields are considered in this thesis, i.e., intelligent transportation system (ITS) and smart grid. Based on the experimental results, a number of optimized schemes on IoT technologies are proposed to improve the reliability and effectiveness of IoT systems. By virtue of ACP approach, virtual-real interaction and parallel control between the artificial systems and the actual systems can be achieved. The contributions of this paper are summarized as follows.
1. For wireless sensor network (WSN) in urban transportation system, we propose the optimized parameter selection methods of low power wide area network (LPWAN) technology for V2X communications. In high-speed scenarios of V2X (vehicle-to-everything), Doppler effect results in time selective fading, which seriously deteriorates the reliability of communication systems. We use ACP-based parallel intelligence theory to solve the time selective fading problem in V2X communication system. An artificial system of V2X communications in urban traffic is constructed, including six vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication scenarios. Computational experiments are executed to analyze the performances of multiple parameter configurations of LPWAN in each V2X scenario. Based on the results of these experiments, the rules of actual V2X communication systems are further extracted. In this way, we design the corresponding optimized parameter schemes for different V2X scenarios. The proposed schemes are capable of combating time selective fading and improving the reliability of systems. Experiments indicate that the proposed schemes achieve better BER performance than traditional schemes.
2. For complex high-speed mobile scenarios, we propose a number of optimized schemes of vehicular multi-carrier communication systems. Orthogonal frequency division multiplexing (OFDM) system is one of the most popular multi-carrier communication systems in vehicular networks. However, in mobile channels, frequency selective fading and Doppler effect are very common due to the high-speed movements of vehicles, which seriously deteriorate the performance of OFDM system, and even lead to inter-carrier interference (ICI) problems. In this thesis, ACP approach is utilized to tackle the aforementioned problems in vehicular networks. We build an artificial system based on IEEE 802.11p to study multi-carrier communication systems. To solve the problem of frequency selective fading in vehicular channels, an interleaved index modulated OFDM (IM-OFDM) is proposed. By optimizing the traditional subcarrier grouping method in IM-OFDM, the interleaved grouping method is able to benefit from the frequency diversity, thus combating frequency selective fading effectively. To solve the problem of Doppler-induced ICI, we propose a novel ICI self-cancellation scheme, which integrates the ICI self-cancellation technique into IM-OFDM framework. The proposed ICI cancellation scheme not only inherits the natural advantage of IM-OFDM, but also suppresses ICI effectively. Experiments indicate that the two proposed schemes achieve better BER performance than other traditional OFDM schemes in vehicular networks.
3. For the field of electric power IoT, we apply narrow-band IoT (NB-IoT) to smart grid and analyze its technical advantages and performance metrics. Due to mission-critical applications, communication technologies with high reliability are required for the smart grid. NB-IoT is able to provide the cellular-level quality of service (QoS), which fully satisfy the reliability requirement in smart grid. NB-IoT also provides wide coverage, low power consumption and massive connections. Thus, NB-IoT is a good choice for communication in the smart grid. In this thesis, NB-IoT is first introduced into the smart gird. The performance comparisons between NB-IoT and other traditional communication technologies are presented. The overall requirements of communications in the smart grid from both quantitative and qualitative perspectives are comprehensively investigated and each of them is carefully examined for NB-IoT. Moreover, based on ACP approach and parallel intelligence theory, we conduct an artificial system based on COST-207 channel model, which contains many typical scenarios in the smart gird, such as urban, rural and hilly terrain area. Computational experiments are conducted on the NB-IoT in various scenarios, which demonstrated that NB-IoT achieves excellent performance in each typical scenario.
|李玉珂. 基于平行智能的物联网关键技术研究及应用[D]. 中科院自动化研究所. 中国科学院大学,2019.|
|Files in This Item:|
|Thesis.pdf（3907KB）||学位论文||开放获取||CC BY-NC-SA||Application Full Text|
|Recommend this item|
|Export to Endnote|
|Similar articles in Google Scholar|
|Similar articles in Baidu academic|
|Similar articles in Bing Scholar|
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.