Efficient sigmoid function for neural networks based FPGA design

Abstract

Efficient design of sigmoid function for neural networks based FPGA is presented. Employing the hybrid CORDIC algorithm, the sigmoid function is described with VHDL in register transfer level. In order to enhance the efficiency and accuracy of implementation on Altera's FPGA, the technology of pipeline and look-up table have been utilized. Through comparing the results obtained by the post-simulation of EDA tools with the results directly accounted by Matlab, it can be concluded that the designed model works accurately and efficiently. © Springer-Verlag Berlin Heidelberg 2006.