Efficient FPGA implementation of Gaussian noise generator for communication channel emulation

In this paper, a high accuracy Gaussian noise generator emulator is defined and optimized for hardware implementation on a FPGA. The proposed emulator is based on the Box-Muller method implemented by using ROM tabulation and a random memory access. By means of accumulations, the central limit method is applied to the Box-Muller output Gaussian distribution. After presenting the algorithmic method, this paper analyzes its efficiency for different noise signal formats. Then the architecture to fit into a FPGA is explained. Finally, results from the FPGA synthesis are given to show the value of this method for FPGA implementation.