Design And Implementation Of Cosine-Modulated Analysis Filter Banks

Multi-rate filter banks is a pop research direction of digital signal processing. It has already applied in several fields, such as audio coding, image compressing, wavelet analysis and so on. As the development of communication technology, the CDMA and OFDM/OQAM multicarrier modulation techniques based on filter banks also become pop research projects.As far as the multi-rate filter banks theory been concerned, how to implement high-speed, efficient filter banks with perfect reconstruction is the critical problem. The principle of polyphase analysis was applied in filter banks design by Vaidyanathan, this simplified the filter banks design in great degree. Koipilai R.D. gave the cosine-modulated filter banks (CMFB) algorithm, in which discrete cosine transform (DCT) was applied to simplify the filter banks structure, which makes the implementation of filter banks in hardware more suitable.FPGA (Field Programmable Gate Array) is the best hardware to implement filter banks for its structure and real-time processing requirement. As the development of EDA (Electronic Design Automation) technology, the method of designing a DSP (digital signal processing) system based on FPGA has improved greatly. Several EDA tools combined makes the designer implement a complicated system in FPGA chips if only he knows the algorithm of system rather than the hardware description language (HDL) and the structure or work principle of FPGA chip. In this thesis, MATLAB, DSP Builder and Quartus Ⅱ are combined to design a cosine-modulated filter banks, and they are key tools for design DSP system.This thesis starts from the developing history of filter banks theory, then the basic principles of filter banks are discussed, including the spectrum change because of decimating or interposing, polyphase and noble equivalent transform. Then the principle...