Seminar: The Jacob Ziv Communication and Information Theory seminar
A Broadcast-Multiple Access Duality in Gaussian Diamond Channels with Inter-Relay Cooperation
| Cloud radio access network (C-RAN) plays a pivotal role in emerging network architectures by facilitating large-scale cooperation among base stations. As the demand for continuously increasing data transmission grows, capacity emerges as the primary bottleneck for every emerging generation of mobile networks. In this context, ultra-dense cell deployment with cooperative operations is anticipated to become a cardinal technology for enabling modern, dependable, ultra-bandwidth, scalable, and rapid communication systems. In contrast to traditional architectures where radio and baseband processing functionality is exclusively implemented within a base station (BS), C-RAN empowers the BS to perform digital processing, digital to analog conversion, analog to digital conversion, power amplification, and filtering, while baseband processing is executed in a central processor (CP) connected to multiple BSs via finite capacity fronthaul and backhaul links. The distinctive characteristics of this architecture render C-RAN capable of managing intensive inter-cell interference in future ultra-dense, multi-tier networks.
In this talk, we illustrate an uplink-downlink duality between the Gaussian diamond BC and the Gaussian diamond MAC with two cooperating relays when compression-based relay coding schemes are employed over the rate-constrained fronthall and cooperation links. Specifically, we investigate compress-and-forward strategies with independent compression across the relays. The oblivious compress-and-forward-based schemes are prominent candidates for practical implementations due to two primary reasons. Firstly, the relays are codebook ignorant, thereby reducing their design complexity. Secondly, efficient implementation utilizing scalar and vector quantizers can approach the theoretical performance. Furthermore, unlike decode-and-forward schemes, which are highly susceptible to poor channel conditions, compression-based approaches can attain theoretical performance bounds under specific conditions. based on joint studies with Shirin Saeedi Bidokhti and Shlomo Shamai (Shitz).
|
| Michael Dikshtein received the B.Sc., M.Sc., and Ph.D. degrees in electrical and computer engineering from Technion, Israel, in 2012, 2019, and 2023, respectively. From 2010 to 2015, he was an RF engineer with Rafael Advanced Defense Systems Ltd. From 2017 to 2018, he was an Automotive Radar researcher at General Motors. He is currently a Bluetooth PHY algorithms researcher at Apple Inc.
|