A Survey of FPGA-Based Hardware Implementation of ANNs

Abstract

In this paper, the development of FPGA-based ANNs is presented. Field programmable gate arrays (FPGA) based artificial neural network (ANN) is now becoming a focus of ANN research. According to the parallelism and nature features of neural network (NN), the hardware implementation is superior comparing with software approach because it can take advantage of these characteristics. Furthermore, since FPGA is a digital device that owns reprogrammable properties and robust flexibility, many researchers have made great efforts on the realization of NN using FPGA technique. However, they encountered many problems during the process. Results of the researches are presented and discussed. The future development of FPGA implementation of ANN is also prospected. Firstly, the introduction of ANN and FPGA technique is briefly shown in this paper. Then ANN implementation with hardware, mainly FPGA, is presented and discussed