In-Cloud Signal Processing and Optimization for 5G Wireless Systems

该提案描述了一个全面的5年研究计划，旨在实现第五代（5G）蜂窝系统的关键方面。5G目前正在加紧开发，以应对当前第四代蜂窝系统不可否认的压力。无线行业为5G设定了非常激进的性能目标，其中包括将关键数据速率指标相对于4G提高100倍或更高。实现这些目标需要从头开始重新思考4G蜂窝架构的各个方面。该研究计划的长期目标是开发分布式优化和信号处理的通用框架，作为未来（5G及以后）可扩展和可持续的蜂窝系统的使能技术。Cloud-RAN是一项颠覆性技术，可以彻底改变我们部署和管理蜂窝基础设施的方式，该计划的短期目标是：进一步开发和分析申请人最近提出的分布式优化和信号处理框架，这对Cloud-RAN.2至关重要。为5G无线通信网络物理层设计高效的大规模分布式信号处理方法。3. 利用机器学习和深度神经网络的最新发展，近实时地解决上述问题。拟议的研究计划将在信号处理、优化、无线通信和机器学习领域产生重大影响的理论和实践进展。重要的是，它将有助于发展对环境负责、可持续和较少干扰的无线基础设施。最后，相关培训计划将为3名博士和4名硕士学生提供上述领域最先进的方法和工具的培训。

This proposal describes a comprehensive 5-year research program which aims at enabling key aspects of the 5th generation (5G) of cellular systems.5G is currently under intense development, in response to the undeniable strain put on the current, 4th generation, cellular systems. The wireless industry has set very aggressive performance targets for 5G that include 100-fold, or higher, increases to key data rate metrics relative to 4G. Attaining these targets requires a rethinking, from the ground up, of all aspects of the 4G cellular architecture.The long term objective of this research program is to develop a common framework for distributed optimization and signal processing as the enabling technology of future (5G and beyond) scalable and sustainable cellular systems. Focusing on Cloud-RAN, a disruptive technology that can revolutionize the way we deploy and manage cellular infrastructure, the short term objectives of the proposed research program are to:1. Further develop and analyze a distributed optimization and signal processing framework that was recently proposed by the applicant, along directions crucial to Cloud-RAN.2. Design efficient large scale distributed signal processing methods for the physical layer of 5G wireless communication networks. 3. Use recent developments in Machine Learning and Deep Neural Networks to solve the above problems in near real-time.The proposed research program will produce both theoretical and practical advances with significant impact in the areas of signal processing, optimization, wireless communications and machine learning. Importantly, it will contribute to the development of environmentally responsible, sustainable and less intrusive wireless infrastructure.Finally, the associated training program will provide training to 3 PhD and 4 Master's student in state of art methods and tools in the aforementioned areas.