BYR Achieve · 镜像论坛

今年是IEEE成立125周年纪念年，IEEE组织了IEEE杰出讲师系列讲座计划（DLP lecture Distinguished Lecturer Program）。应IEEE电路与系统北京分会与光通信与光电子学研究院、IEEE高级会员张晓光教授邀请，美国明尼苏达大学教授、IEEE杰出讲师G..Sobelman博士（中文名字苏杰睿）于6月12日来北邮进行IEEE杰出讲师系列讲座计划之一——北京邮电大学IEEE杰出讲师讲座。讲演后，苏教授将与北邮IEEE会员、IEEE学生会员进行座谈。报告题目：Flexible VLSI Architectures for UWB, MIMO and Cognitive Radio（超宽带、MIMO系统与无线识别的大规模集成电路结构）主讲人：Prof. Gerald E. Sobelman （Dept. of Electrical and Computer Engineering University of Minnesota）时间：2009年6月12日（星期五）下午14:00-16:00 地点：教三楼136报告厅内容：Flexibility and efficiency will be key design requirements for future communications transceivers. Systems must be reconfigurable and adaptable in order to minimize power and maximize bandwidth as environmental conditions and user requirements change. This lecture will present architectures and design techniques to achieve these goals for Ultra Wideband (UWB), multiple antenna (MIMO) and Cognitive Radio systems. We will begin with a class of UWB systems based on Orthogonal Frequency Division Multiplexing (OFDM). The standard approach uses large FFTs, which can require a significant amount of hardware resources. An attractive design alternative known as Pulsed-OFDM UWB can achieve the same system-level performance while requiring less hardware and power. We will describe flexible architectures to implement these OFDM-based systems. In addition, we will also present a hardware model of the UWB channel in order to achieve faster system-level verification. The performance of MIMO systems scales with the number of antennas, so extremely high throuphputs are possible. In practice, however, the design complexity also increases rapidly with the number of antennas, so it is crucial to develop efficient implementations for such systems. We will describe a promising approach based on the Geometric Mean Decomposition (GMD) and we will give throughput and design complexity results for a range of MIMO systems. Cognitive radio seeks to take advantage of the local spectrum conditions in order to maximize throughput whenever possible. This requires a flexible platform in which many aspects of the system can be reconfigured dynamically. For example, the parameters associated with digital filters, FFTs and interleavers may have to be adjusted in order to implement a new data rate, modulation scheme or detection algorithm. In addition, the connections between various functional units may have to be altered to implement different data flow requirements. Programmable functional units and network-on-chip architectures will be described to address these issues.

[讲座]IEEE杰出讲师系列讲座（转载）