Integrated Waveform and Intelligence (IWAI): Physical Layer Solutions to Sustainable 6G

通信和信息网络在我们的日常生活中无处不在，据估计，信息和通信技术（ICT）占碳排放量的1.8%-2.8%。5G的成功部署表明，由于5G的基站更密集，5G通信系统的总功耗远高于4G。最近的一项研究还显示，由于天线数量更多，频谱带宽占用更宽，基站部署策略更密集，潜在的6G系统消耗的功率将比5G系统高近50倍。5G技术已经标准化并在现实生活中部署，但没有优先考虑净零可持续性。因此，为了引领下一代零净通信的可持续性创新，我们应该重塑6G的物理层技术。提高能源效率已成为全球的优先事项，并已被确定为下一代6G系统的关键技术推动者，许多行业和研究机构正在追求净零可持续技术，以实现减少能源使用和碳排放的目标。该项目旨在支持可持续通信系统的研究，这与英国政府在《无线基础设施战略：2030年愿景》中的首要任务和净零目标相一致（来源：GOV.UK）。通信系统的功耗与物理硬件、物理信号等物理部件直接相关。通过先进制造进行硬件升级可以降低功耗，但当需要更多基站为特定区域提供服务时，6G的贡献有限。因此，高效节能物理信号设计的根本性突破，更精确的波形设计，是及时和专门定位的，以实现6G的净零目标。自1924年哈里·奈奎斯特（Harry Nyquist）开发当今数字通信信号的基础以来，已经取得了许多进步。然而，现有的通信系统波形设计面临着许多根本性的挑战(a)现有的节能空中接口波形设计受到Mazo极限的限制，在没有任何性能损失的情况下只能节省高达20%的功率。然而，20%的功耗节省已经不够了，20%的功耗是通过复杂和耗能的信号处理来实现的。(b)数学推导的波形受已知数学模型的限制，不太可能是最佳的。(c)机器学习可以帮助信号具有更好的性能，但机器学习算法需要大量的处理能力。(d)由于需要强大的计算资源，现有的信号波形和人工智能模型通常使用高能耗硬件进行部署。为了应对上述挑战，提出了一个雄心勃勃的计划，包括i)超越传统20%节电限制的新波形的基础探索，ii)波形和人工智能的共同设计，以获得先进的波形格式并降低人工智能模型的复杂性，iii)具有节电验证的低成本硬件概念验证。具有智能的信号波形设计的基础研究突破将影响从电路设计到新型通信系统、人工智能算法、传感器和传感器网络、绿色通信技术、先进信号处理、生物医学信号等许多领域和领域的研究人员的工作。

Communications and information networks are ubiquitous in our daily life, and information and communications technology (ICT) is estimated to contribute 1.8%-2.8% of carbon emissions. The successful deployment of 5G has revealed that the total power consumption for a 5G communication system is much higher than that in 4G due to more densely placed base stations in 5G. A recent study also revealed that a potential 6G system would consume nearly 50 times higher power than a 5G system due to a larger number of antennas, a wider spectral bandwidth occupation, and a denser base station deployment strategy. 5G techniques have already been standardized and deployed in real life without the priority on net-zero sustainability. Therefore, to lead the sustainability innovations for next generation net-zero communications, we should reshape the physical layer techniques for 6G.Improving energy efficiency has become the priority worldwide and has been identified as a key technology enabler for next generation 6G systems, and a number of industries and research institutes are chasing for net zero sustainable technology to achieve the objectives of cutting energy usage and carbon emission. This project aims for underpinning research in sustainable communication systems, which aligns with the UK government's top priority and ambitions on net zero in Wireless Infrastructure Strategy: a vision for 2030 (Source: GOV.UK). Power consumption for a communication system is directly linked to physical components such as physical hardware and physical signals. Hardware upgrade via advanced manufacturing can cut power consumption but with limited contributions in 6G when more base stations are required to serve a given area. Therefore, a fundamental breakthrough in energy efficient physical signal design, more precisely waveform design, is timely and specially positioned to achieve net-zero goals in 6G.A number of advancements have been achieved since 1924 when Harry Nyquist developed the foundation of today's digital communication signals. However, the existing waveform design in communication systems faces a number of fundamental challenges (a) Existing energy efficient air interface waveform designs are limited by the Mazo limit, which can only save power by up to 20% without any performance loss. However, the 20% power saving is no longer sufficient and the 20% is achievable at the cost of sophisticated and energy consuming signal processing. (b) Mathematically derived waveforms are limited by known mathematical models and are unlikely to be the optimal. (c) Machine learning can assist signals to have better performance but machine learning algorithms require lots of processing power. (d) Existing signal waveforms and AI models are commonly deployed using high energy consuming hardware because powerful computing resources are needed. To address the above challenges, an ambitious program is proposed including i) fundamental explorations of new waveforms beyond the conventional 20% power saving limit, ii) co-design of waveform and artificial intelligence to derive advanced waveform formats and cut the complexity of AI models, iii) low-cost hardware proof of concept with power saving validations.The fundamental research breakthrough in the signal waveform design with intelligence will have impacts on researchers' work across many areas and fields, from circuit design to new communication systems, artificial intelligence algorithms, sensors and sensor networks, green communication techniques, advanced signal processing, biomedical signals, and other areas.