#finalyearprojects

Fast Fourier Transform (FFT) is a numerical algorithm which widely used in digital signal processing for the analysis of the signal in frequency-domain. Although FFT IP cores have been developed by various companies for general utilization in various applications, most of these IP cores have many features that sometimes are redundant for a specific application. The project aims to implement FFT algorithm in hardware with negligible error in converting the function with time-domain representation to the function with frequency-domain representation. In this project, the FFT algorithm pseudocode is designed in C language first before implementing it in Verilog language so that the functionality of the algorithm can be verified with the FFT built-in function in MATLAB. To reduce hardware complexity, the Q15.16 fixed-point number system is used instead of floating-point. Error analysis of Verilog and C implementation shows an average percentage difference of 0.98% compared to the gold standard in MATLAB. This corresponding to less than 0.02 dB error. In conclusion, the FFT algorithm designed in Verilog language able to convert the discrete-time signal to the discrete-frequency signal since it has the same functionality as the FFT function in MATLAB.

Contact supervisor: Muhammad Mun'im bin Ahmad Zabidi
Department of Electronics and Computer Engineering
School of Electrical Engineering
Faculty 0f Engineering
Universiti Teknologi Malaysia
Johor Bahru, Malaysia.
[email protected]

FastFourierTransform