题目：Multi-Hop Wireless Pipeline Powering Internet-of-Things 时间：2018年7月9日（周一）上午10:00-11:30 地点：信电大楼-215学术厅 报告人：Prof. Lin Cai, University of Victoria

报告摘要: Connecting more than half of world population, the Internet changes how we acquire, generate and exchange information and knowledge, and how we work, live, and relate to one another, forever. Enabled by advanced wireless technologies, the Internet will connect billions to trillions of devices in the coming decades, leading to a new era, Internet-of-Things (IoT). Wireless transmission between two devices  is a mature technology. However, long-distance wireless transmission for IoT is technically challenging. Wireless communication is broadcast in nature, so signals are propagated and decayed over distance and interfere with each other. To enable long-distance information exchange, it is sensible to construct a multi-hop bidirectional wireless pipeline, so wireless devices can use low power with less interference to relay for each other, to/from the Internet, achieving a higher spatial capacity. However, multi-hop wireless transmission has suffered from the curse that the throughput and reliability degrade quickly when the number of hops increases, due to mutual interference and inefficient resource management. Realizing that the layered Internet architecture designed for wired networks does not fit well for wireless ones, we abandoned the traditional approach that the physical (PHY) layer deals with point-to-point transmission only, and intentionally let the PHY layer be adaptive to network topology and traffic. Following this line of breakthrough, we proposed a multi-hop physical-layer network coding (MPNC) solution to create a wireless bidirectional pipeline that can maintain an end-to-end (e2e) scalable throughput regardless of the distance, i.e., the throughput of an any-hop path is the same as that of a one-hop transmission if the per-hop distance remains the same, while the e2e bit-error-rate increases merely linearly w.r.t. hop count using an effective error cancellation design. To prove the concept of this non-traditional approach, we have built a software-defined radio testbed to implement the MPNC solution. The test results confirm that, in a static environment, MPNC is feasible to achieve the best possible performance for multi-hop wireless transmissions, a major breakthrough towards the holy grail in wireless networking.

蔡霖教授简介：

加拿大维多利亚大学的终身教授蔡霖是无线网络资源管理和性能分析方向的知名学者，在网络跨层优化、资源调度、拓扑控制、物联网和多媒体传输等领域取得了众多成果，在国际知名杂志和会议发表文章 200多篇, 其中IEEE汇刊/Automatica期刊文章80多篇，IEEE/ACM旗舰会议文章90多篇，并著学术专著2本, 分别由John Wiley & Sons出版社和Springer出版社出版。荣获得IEEE旗舰会议IEEE ICC 2008最佳论文奖和IEEE WCNC 2011最佳学术论文奖。鉴于她的杰出学术贡献和原创性研究成果，于2010年和2015年荣获加拿大国家自然科学和工程基金会(NSERC)支持卓越研究的“发现加速器奖”（Discovery Accelerator Supplements Award)。蔡霖教授(曾)担任多个知名期刊的编委, 包括IEEE Internet of Things Journal, IEEE Trans. on Wireless Communications和IEEE Trans. on Vehicular Technology等， 2018年起担任大数据领域重要期刊IEEE Trans. on Big Data的指导委员会(Steering Committee)成员。她创立了加拿大维多利亚地区IEEE车辆技术和通信联合协会(VTS/ComSoc)并担任主席职务，2018年起任加拿大国家自然科学和工程基金会Discovery Grant的专家评委，是BC省注册专业工程师。