Future Millimetre Wave RF Transceiver Architectures for Communications Systems

6千兆赫以下的无线电频谱不足以应付未来的流动通讯需求。研究人员现在也开始考虑2030年智能信息社会的需求，这可能包括进一步推进亚太赫兹无线电频谱，以提供更多的用户数据带宽。在5G、未来6G及以后，在固定无线接入和手持设备中使用毫米波（mmWave）频段将需要节能、低成本且高性能的射频收发器。这种收发器还必须支持极高的数据速率（例如，千兆以太网；USB 3.0的5Gbit/s；车载“信息娱乐”的10gbit /s峰值速率和用于虚拟现实的8k超高清电视）。迄今为止，这一具有挑战性的要求在所有传统毫米波技术中都是相互排斥的。随着搭载28GHz / 39GHz通信无线电（n257 ~ n261频段）的三星电子“盖乐世S10 5G”等早期5G智能手机的上市，毫米波移动通信时代开始了。尽管入门级5G处于早期部署阶段（使用对4G的修改），但在2030年之前，它不太可能被定义或在高毫米波频段（约73GHz）部署。初步分析表明，多gbit /s数据所需的数字信号处理（DSP）可能会扩展到每秒100亿条“乘法累积”指令。当与模拟无线电功能结合使用时，仅接收功能就会消耗14W的电池功率，而发送功能则会消耗更多的电池功率。目前，智能手机的电池容量约为4.5Ahr，在这种消耗的接收功率下，只能支持1小时的运行。因此，迫切需要对E波段（约73GHz）及以上频率的毫米波无线电硬件和软件架构进行新的研究。该奖学金将重点关注以下领域：1)形成毫米波天线阵列的成本效益和节能技术。我们最近对时间调制天线阵列（TMA）的研究显示了在较低频率下提高TMA效率的方法。TMA的一个主要吸引力是其简单的控制界面（全数字）。2)重新研究基本毫米波电路概念，如混频器和振荡器，使用新的见解和利用最新的发现来制造关键部件，如谐振器。毫米波谐振器的研究现在可以受益于谢菲尔德大学最新的3D打印技术以及低温共烧陶瓷的最新技术。3)毫米波收发器在硬件和软件方面的整体视图，并通过分区为射频性能目标提供最佳功率效率。在具有数百个天线和收发器的大规模多输入多输出系统（M-MIMO）中，新技术将在节省功耗方面发挥重要作用。在现有的M-MIMO系统中，射频硬件消耗的功率可以与数字信号处理器相媲美。研究将包括重新考虑长期被遗忘的电路拓扑和思想，在这个新的设置。4)探索毫米波认知无线电的信号处理技术-使其能够感知其操作环境并优化其性能（通过可重构射频硬件）。此外，人工智能的出现和能力的提高现在与更接近硬件本身的操作相关，例如解调传入的RF信号。5)项目期间将创建原型测试芯片和子系统。这些将用于建立毫米波无线电系统演示，包括传播测量研究。试验平台的奖学金后申请将支持未来大规模生产毫米波系统的进一步研究，并促进增强的行业参与。

There is insufficient radio spectrum below 6GHz to cater for future mobile communications demand. Researchers are also now beginning to consider the needs of the 2030 intelligent information society, which will likely include a further push into sub-terahertz radio spectrum, to deliver yet more user data bandwidth.In 5G, future 6G and beyond, use of millimetre wavelength (mmWave) bands in fixed wireless access and handheld equipment will require power efficient, low cost yet high-performance RF transceivers. Such transceivers must also support extremely high data rates (e.g. Gigabit Ethernet; 5Gbit/s for USB 3.0; 10's of Gbit/s peak rates for vehicular 'infotainment' and '8k' ultra-high-definition TV for virtual reality). This challenging set of requirements has, to date, been mutually exclusive in all conventional mmWave technologies.With the release of early 5G smartphones, such as Samsung Galaxy S10 5G incorporating 28GHz / 39GHz communication radios (bands n257-n261), the era of mmWave mobile communications has begun. Although entry-level 5G is in early stage deployment (using modifications to 4G), it is unlikely to be defined or viable for deploying at high mmWave bands (circa 73GHz) before 2030.Initial analysis shows the digital signal processing (DSP) required for multi-Gbit/s data may extend to 10's of billions of 'multiply-accumulate' instructions per second. When combined with analogue radio functions, this could result in consumed battery powers of 14W by receive functions alone, with considerably more for transmit. Smartphone battery capacities are now circa 4.5Ahr, which would support just 1 hour of operation at such consumed receive powers.Thus, there is an urgent need for new research into mmWave radio hardware and software architectures, for frequencies at E band (circa 73GHz) and beyond.The Fellowship will focus on the following areas:-1) Cost-effective and power-efficient techniques to form mmWave antenna arrays. Our recent research into Time Modulated Antenna Arrays (TMA) has shown ways of improving TMA efficiency at lower frequencies. A key attraction of the TMA is its simplicity of control interface (all digital).2) Reinvestigation of fundamental mmWave circuit concepts, such as mixers and oscillators, using new insight and making use of the latest findings for manufacturing key components such as resonators. The research in resonators at mmWave could now benefit from the latest 3D printing techniques available at the University of Sheffield as well as updated techniques in low temperature co-fired ceramics. 3) A holistic view of the mmWave transceiver in terms of hardware and software, with partitioning to give best power efficiency for an RF performance target. Novel techniques will be valuable in saving power in massive multiple-input multiple-output systems (M-MIMO), having many hundreds of antennas and transceivers. In existing M-MIMO systems the power consumed by RF hardware could rival that of the digital signal processors. Research will include reconsidering long-forgotten circuit topologies and ideas, in this new setting. 4) Exploration of signal processing techniques for mmWave cognitive radio- allowing it to sense its operational environment and optimise its performance (via reconfigurable RF hardware). Also, the emergence and increase in capability of artificial intelligence is now becoming relevant for operation closer to the hardware itself, such as in demodulating an incoming RF signal.5) Prototype test chips and subsystems will be created during the project. These will be used to build mmWave radio system demonstrators, including for propagation measurement research. The post-fellowship application for the trial platforms will support further research in future mass-producible mmWave systems, as well as facilitating enhanced industry engagement.