Application of CORDIC Algorithm to Neural Networks VLSI Design

Application of coordinate rotation digital computer (CORDIC) algorithm to neural networks VLSI design is illustrated in this paper. The hybrid CORDIC algorithm, which combines the basic CORDIC with LUT, is introduced to design the net function and activate function of the neural networks. Take the trigonometric function for example, the procedure of implementation on FPGA based hybrid CORDIC is proposed, by which other function computations can be easily designed and realization. The simulation and synthesis results show that the designed RTL model coded in VHDL has the characteristic of accuracy and efficiency

[1]  Earl E. Swartzlander,et al.  Hybrid CORDIC Algorithms , 1997, IEEE Trans. Computers.

[2]  Martine Wedlake,et al.  DIGITAL ARTIFICIAL NEURON , 1997 .

[3]  J. S. Walther,et al.  A unified algorithm for elementary functions , 1899, AFIPS '71 (Spring).

[4]  M. Wedlake,et al.  A CORDIC implementation of a digital artificial neuron , 1997, 1997 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, PACRIM. 10 Years Networking the Pacific Rim, 1987-1997.

[5]  Uwe Meyer-Baese,et al.  Digital Signal Processing with Field Programmable Gate Arrays , 2001 .

[6]  Jack E. Volder The CORDIC Trigonometric Computing Technique , 1959, IRE Trans. Electron. Comput..