CAREER: Error-control for Streaming Media: Architecture, Code Design, and Noisy Feedback

摘要：“该奖项是根据2009年美国复苏和再投资法案(公法111-5)资助的。”数字通信技术是建立和加强全球在线社区的核心。数据网络越来越多地为语音、视频、控制和交互应用程序传输延迟敏感的流内容；这些数据具有网络传统上未针对其进行优化的特性。该项目旨在重新思考数字通信的基本架构，以改进流数据的交付。长期的潜力是以更低的成本改善服务，并扩大可获得性。核心技术问题是，在面临资源受限和噪声反馈的真实系统中，如何最好地利用反馈来传输延迟敏感数据。虽然反馈是稳健系统设计的一项基本原则，但在确保可靠通信方面的使用仅限于现场视察堆栈的较高层。本项目探讨了通过利用物理层的反馈实现的性能改进。首席调查员(PI)已经演示了如何通过转移到错误和擦除的角度来利用噪声反馈，在反向链路上利用识别码，并新颖地使用不等错误保护。其结果是大大降低了误码率和延迟。该项目将针对新的场景(例如，无线衰落、多用户网络、与排队的交互)来开发体系结构洞察力、适当的纠错码类别和改进的网络协议。这项工作将被整合到华盛顿大学麦迪逊分校的课程中，包括编码理论、信息论和本科生课程“社会概论？S工程重大挑战”。PI和他的研究团队还通过威斯康星大学工程暑期项目参与了高中STEM的推广活动。将通过皮奥？S现有的产业合作加速将其转化为实践。

Abstract:"This award is funded under the American Recovery and Reinvestment Act of 2009(Public Law 111-5)."Digital communication technology is central to the establishment and enhancement of global on-line communities. Increasingly, data networks convey delay-sensitive streaming content for voice, video, control, and interactive applications; data with characteristics for which networks have not traditionally been optimized. This project aims to rethink the fundamental architectures of digital communications to improve delivery of streaming data. The long-term potential is improved service at reduced cost, and broadened availability.The central technical question is how feedback should best be used to transmit delay-sensitive data in real-world systems that are faced with resource-constrained and noisy feedback. While a fundamental principle of robust system design, the use of feedback in ensuring reliable communication has been limited to higher layers of the OSI stack. This project explores the performance improvements made possible by taking advantage of feedback at the physical-layer. The principal investigator (PI) has demonstrated how to exploit noisy feedback by shifting to an error-and-erasures perspective, utilizing identification codes on the reverse link, and making novel use of unequal error protection. The result is a massive reduction in bit-error rate and delay. This project will address new scenarios (e.g., wireless fading, multiuser networks, interaction with queuing) to develop architectural insights, appropriate classes of error correcting codes, and improved network protocols.This work will be incorporated into courses at UW Madison including coding theory, information theory, and the undergraduate course "Introduction to Society?s Engineering Grand Challenges". The PI and his research team are also involved in high-school STEM outreach through the UW Engineering Summer Program. Translation to practice will be hastened through the PI?s existing industrial collaborations.