Implementation of 64-point FFTs as a building block for OFDM-based standards with respect to different criteria

• Main Author: Krishnaih Lokanadhan, Sarojanarayanamurthy Rajavari
• Format: Others
• Language: English
• Published: KTH, Skolan för informations- och kommunikationsteknik (ICT) 2013
• Subjects: Computer and Information Sciences; Data- och informationsvetenskap
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-140885

Abstract This master thesis implements a critical block: Fast Fourier Transform (FFT)/Inverse Fast Fourier Transform (IFFT) for standards using Orthogonal Frequency Division Multiplexing (OFDM). They are used in OFDM systems for modulating / demodulating subcarriers. The main goal of this thesis is to implement these blocks in a Field Programmable Gate Array (FPGA) using Very High Speed Integrated Circuit Hardware Description Language (VHDL). OFDM technology is used in multiple wireless communication standards including WLAN IEEE 802.11(a/g/n), WiMAX, LTE and DVB. 64-point FFT/IFFT blocks determines the complexity of an IEEE 802.11 OFDM transceiver. In this thesis project, a serial to parallel (S2P) converter, six FFT/IFFT architectures and parallel to serial (P2S) converter modules using fixed-point two's complement number system were implemented. The FFT uses an 18 bit fixed point representation for input/output data and memory coefficients. It is designed using radix-2 Decimation-In-Time (DIT) architecture. The implemented FFT architectures are differing with respect to RAM memory access, way of processing of the input samples, number of computational elements (butterflies) and memory storage requirements. The three realized implementation categories are: in-place computation, direct FFT computation and pipelined FFT computation. This report explains FFT architectures and methodologies to design, simulate and implement them on an FPGA device. This thesis report also presents area results in terms of FPGA resources and throughput results in terms of number of clock cycles for all the FFT architectures.