Performance Assessment of Model-Driven FPGA-based Software-Defined Radio Development

• Main Author: Allen, Matthew S
• Other Authors: Alexander M. Wyglinski, Advisor
• Format: Others
• Published: Digital WPI 2014
• Subjects: HDL Coder; SDR; FPGA
• Online Access: https://digitalcommons.wpi.edu/etd-theses/943; https://digitalcommons.wpi.edu/cgi/viewcontent.cgi?article=1942&context=etd-theses

"This thesis presents technologies that integrate field programmable gate arrays (FPGAs), model-driven design tools, and software-defined radios (SDRs). Specifically, an assessment of current state-of-the-art practices applying model-driven development techniques targeting SDR systems is conducted. FPGAs have become increasingly versatile computing devices due to their size and resource enhancements, advanced core generation, partial reconfigurability, and system-on-a-chip (SoC) implementations. Although FPGAs possess relatively better performance per watt when compared to central processing units (CPUs) or graphics processing units (GPUs), FPGAs have been avoided due to long development cycles and higher implementation costs due to significant learning curves and low levels of abstraction associated with the hardware description languages (HDLs). This thesis conducts a performance assessment of SDR designs using both a model-driven design approach developed with Mathworks HDL Coder and a hand-optimized design approach created from the model-driven VHDL. Each design was implemented on the FPGA fabric of a Zynq-7000 SoC, using a Zedboard evaluation platform for hardware verification. Furthermore, a set of guidelines and best practices for applying model-driven design techniques toward the development of SDR systems using HDL Coder is presented."