FPGA-Based Real-Time Signal Processing Architecture for High-Speed Communication Systems
Keywords:
FPGA, Real-time signal processing, High-speed communication systems, Parallel architecture, Low-latency design, High-throughput processingAbstract
The growing need to have high-speed communication protocols and systems like 5G/6G networks, software-defined radio (SDR), and real-time signal processing applications, requires architectures that can provide low latency and high throughput. The traditional processor-based solutions such as the CPUs and GPUs tend to not handle the strict real time requirements because of their sequential processing and consuming large amounts of power. The paper describes a real-time signal processing architecture based on the FPGA that is aimed at providing high-throughput and low-latency to highspeed communication systems. The architecture proposed uses parallel processing, pipelined data flow and the effective use of on-chip resources like DSP slices and block RAM to optimize the use of the on-chip resource. A mathematical model is built to study system throughput and latency, and inform the architectural design. The implementation is on a Xilinx Zynq-7000 FPGA platform and tested at high data-rate conditions. Experimental findings show a high processing throughput and latency reduction over traditional CPU- and GPU-based implementations at a low power consumption. The system proposed provides a scalable, low energy solution that can be used in the next generation of real time communication systems

