Architectural Advantages, Applications, and Future Trends of FPGAs in Real-Time Edge Computing: A Review

Abstract

Edge computing is with the continuous development of real-time sensing applications, artificial intelligence, Internet of Things, etc. In order to speed up data processing and reduce waiting time, FPGAs, as the base of hardware that can change its structure again, has great advantages in real-time edge computing thanks to its parallel processing ability, data channel with less waiting time, good energy efficiency and easy matching design. Compared with CPU, GPU and ASIC, this paper pays attention to the design advantages and uses of FPGAs in real-time edge computing, which can be processed side by side with less waiting time, energy efficiency and reconfiguration. In addition, we will also check typical edge computing scenarios, such as AI inference, real-time image processing, industrial control and intelligent sensing. Finally, we found the big topic of current research, talked about the new trends, and put forward ideas on the future direction of edge computing design using FPGA.