Ricerca CERCA

5G Active Antenna profiling by using FPGA acceleration


Reference persons LUCIANO LAVAGNO

External reference persons Roberto Quasso, Telecom Italia Wireless Access Innovation, Torino

Research Groups High-Level Synthesis and FPGA acceleration


Description Active Antennas are expected to play a fundamental role in LTE/5G Mobile Networks and profiling their performance is very important both in ideal conditions (e.g. in anechoic chamber) and in real condition (when installed in the field). Unlike traditional antennas, which were driven by analog radio signals, active antennas receive a digital optical signal from the mobile station and internally convert it to a radio signal; this means that the measurement equipment must be able to synchronize and analyse the digital content of the radio signal in order to assess the antenna’s performance.
TIM laboratories carry out this activity with both commercial instruments and customized SDR FPGA-based devices, the latter solution providing more flexibility and cost-effectiveness for special use cases.
As part of the TIM activity based on these custom SDR tools, the thesis goal is to enrich the current platform already available for 5G Active Antenna profiling, by defining and developing new functions from the physical layer of the 5G radio standard, to address new use cases and situations.
Specifically, the thesis work will involve first a clock-cycle accurate simulation stage (for evaluation of performance and tradeoffs) and then the implementation stage which could be as FPGA-only blocks or as FPGA blocks tightly coupled with some real-time SW module.

Notes C/C++,Python and Verilog programming experience is mandatory, while OpenCL knowledge is a plus; ability to “think parallel” and manage streaming and concurrent processes will be important; knowledge of FPGA devices and high-level synthesis is very useful. Knowledge of Matlab/Simulink and common DSP functions (FIR, FFT, etc.), which are available as library blocks is useful. Knowledge of radio and antennas may help but is not mandatory.

Deadline 17/07/2020      PROPONI LA TUA CANDIDATURA