Kees Van Berkel, STEricsson, Eindhoven University of Technology, NL


Making Data Flow in 4G Wireless Transceivers


Wireless data rates have increased tenfold every five years for nearly two decades and will continue to do so for at least another five years. The prevailing 4G communication standards WCDMA, LTE-A, and 802.11ac support several hundreds of megabits per second, with signal processing requirements approaching one terra operations per second. For this task, smartphone thermal limits allow a power budget that is well below 1W. Furthermore, some smartphone use cases require multiple transceivers to run simultaneously, where the corresponding standards specify real-time requirements from one millisecond down to the microsecond range. Accordingly, the problem of organizing the 4G wireless signal and protocol processing in smartphones can be summarized as hard real-time multi-tasking on a tightly constrained heterogeneous multi-core architecture. In this presentation we will explore formal dataflow techniques to address this problem, focusing on the central question: is the data flow programming model sufficiently expressive to describe these transceivers accurately and at the same time sufficiently constraining to enable rigorous real-time analysis and efficient resource usage?


Kees van Berkel is a fellow at ST-Ericsson since 2008, after spending most of his R&D career at Philips and NXP Research. Since 1996 he is also a part-time professor in Computing Science at the TU Eindhoven. Kees pioneered asynchronous VLSI from theory to mass production, as well as embedded vector processing for modems and software defined radio. His current research is on multi-standard digital transceivers, vector processors, multi-core architectures, resource management, and low power,

* If you wish to modify any information or update your photo, please contact the web chairalain.fourmigue@polymtl.ca