Coded multi-terminal communication systems: theory, applications and experiments

绝大多数实际通信场景涉及多个终端和/或用户，例如，在无线和光学数据接入网中，在多天线（MIMO）信道上传输，在相关监视图像或传感器测量的分布式压缩中。网络信息理论（多终端系统的信息理论）的结果表明，与基于传统单终端通信技术的通信架构相比，适当的多终端编码可以在传输速率、可压缩性、可实现的频谱效率等方面带来巨大的性能提升。这些增益通常是数量级的，例如，与传统上使用的随机接入网络协议或基于单用户的CDMA传输相比，具有高斯噪声的多接入信道的总吞吐量理论上可以通过多用户解码增加多达20倍。同样，空时编码和MIMO通信系统可以将无线通信链路的容量和频谱效率提高十倍以上。最后，无线传感器网络中有效的分布式压缩只有通过对睡眠狼问题进行适当的网络编码才能实现。实现上述巨大的性能提升通常会在通信系统理论、编码、调制/解调、压缩等方面带来开放的问题和新的未探索的场景。本提案的主要目标是继续建立一个编码多终端系统的研究项目，即涉及多个发射器和/或接收器的通信系统。首席研究员的目标是利用他和他的研究团队在过去5年中在这一领域获得的动力和专业知识。我们已经成功地展示了几个世界第一，例如，睡眠狼问题中的涡轮压缩，多接入信道编码的过载OCDMA网络传输，具有数十个天线但实际复杂性的上行MIMO系统的近容量传输。本提案的具体研究目标和方法可概述如下：1)理论：(a)设计近香农极限实现turbo码和迭代解码算法，选择实际感兴趣的多终端问题；(b)为涡轮编码多终端系统导出分析的封闭式误码率评价技术；(c)使用信息论的应用工具对系统参数进行稳健性和敏感性分析。2)应用：将设计的turbo码，解码算法和/或性能评估技术应用于实际感兴趣的通信问题，例如(a)为3G和4G蜂窝标准设想的基于mimo的无线数据接入网络；(b)光CDMA网络传输过载；(c)具有多磁道和读头的磁和光学记录系统；(d)用于监视数据和无线传感器网络的误差稳健和分布式涡轮压缩技术。3)实验：利用我们最近建立的基于fpga的硬件软件平台进行多终端算法开发和实时测试(a)开发所提出编码技术的实时实现；(b)利用离线信道测量和真实数据源，在实时模拟通信情况下测试这些技术；(c)促成并参与将已开发的技术转让给系统示范和原型设计平台的可能性。就所提出的研究计划的意义而言，全球通信系统正迅速向随时随地可用的高速数字服务发展。每年，这些系统产生数十亿美元的收入，并依赖于特定通信场景的系统级解决方案（无线，光学或记录；语音，数据或视频；上行或下行链路；宏，微或微单元等）预期的理论和应用贡献是编码，调制和多终端通信系统的性能评估。所获得的研究成果有望对全球通信基础设施的低复杂度功率和带宽效率收发器设计产生影响。此外，根据首席研究员过去5年的业绩记录，在拟议项目期间接受密切指导并获得申请资金支持的研究生（和本科生）将受到著名研究实验室和电信公司的需求。

A vast majority of practical communication scenarios involves multiple terminals and/or users, e.g., in wireless and optical data access networks, in transmission over multi-antenna (MIMO) channels, in distributed compression of correlated surveillance images or sensor measurements.  Results in network information theory (information theory of multi-terminal systems) indicate that appropriate multi-terminal coding can result in huge performance gains, in terms of transmission rates, compressibility, achievable spectral efficiencies, etc., when compared to communication architectures based on traditional single-terminal communication techniques. These gains are often of the order of magnitude, e.g., the aggregate throughput of a multi-access channel with Gaussian noise can be theoretically increased by as much as 20 times with multi-user decoding, when compared to traditionally used random access network protocols or single-user based CDMA transmission. Similarly, space-time codes and MIMO communication systems can offer more than ten-fold increases in capacities and spectral efficiencies of wireless communication links. Finally, efficient distributed compression in wireless sensor networks can be only achieved with appropriate network coding for the Slepian-Wolf problem.              Realizing the mentioned large performance gains leads often to open problems and novel unexplored scenarios in communication systems theory, coding, modulation/demodulation, compression, etc. The main objective of this proposal is to continue building a research program in coded multi-terminal systems, i.e., communication systems that involve more than one transmitter and/or receiver. The principal investigator aims to use the momentum and expertise in this area, obtained by him and his research team during the past 5 years. We have successfully demonstrated several world firsts, e.g., turbo compression in the Slepian-Wolf problem, overloaded OCDMA network transmission with multi-access channel coding,  near-capacity transmission for uplink MIMO systems with tens of antennas yet practical complexity.                The specific research objectives and approaches of this proposal can be outlined as follows: 1) THEORY: (a) Design near-Shannon limits achieving turbo codes and iterative decoding algorithms for selected multi-terminal problems of practical interest;  (b) Derive analytical, closed form BER evaluation techniques for turbo coded multi-terminal systems; (c) Perform robustness and sensitivity analysis of system parameters using applied tools from information theory.   2) APPLICATIONS: Apply the designed turbo codes, decoding algorithms and/or performance evaluation techniques to communication problems of practical interest, e.g., (a) MIMO-based wireless data access networks envisioned for the 3G and 4G cellular standards; (b) Overloaded optical CDMA network transmission; (c) Magnetic and optical recording systems with multiple tracks and read-heads; (d) Error robust and distributed turbo compression techniques for surveillance data and wireless sensor networks. 3) EXPERIMENTS: Using our recently set-up FPGA-based hardware-software platform for multi-terminal algorithm development and real-time testing (a) Develop real-time implementations of the proposed coded techniques; (b) Test these techniques in real-time, emulated communication scenarios using off-line channel measurements and real data sources; (c) Enable and participate in potential transfer of developed technologies to system demonstrators and prototyping platforms.             In terms of the significance of the proposed research program, global communication systems are progressing rapidly towards high-speed digital services available anytime and anywhere. Every year, these systems are generating multibillion dollar revenues and rely on system level solutions which are specific to the particular communication scenario (wireless, optical or recording; voice, data or video; uplink or downlink; macro, micro or pico cell, etc.)   Expected theoretical and applied contributions are in coding, modulation and performance evaluation of multi-terminal communication systems. Obtained research results are expected to have impact on the low complexity design of power and bandwidth efficient transceivers for the global communication infrastructure.   In addition, graduate (and undergraduate) students trained in close mentorship during the proposed program and supported via the requested funding will be in demand by prestigious research labs and telecom companies, as indicated by the principal investigator's track record over past 5 years.