Radix-22 based low power reconfigurable FFT processor

Fast Fourier transform (FFT) is widely applied in the speech processing, image processing, and communication system. To implement it, a radix-2² based reconfigurable FFT processor is proposed in this paper. This architecture gets the optimal balance between flexibility and power consumption. Power saving is achieved by using the appropriate FFT size instead of a fixed large FFT size. The memory-based architecture is used to design our reconfigurable FFT processor. It can be configured to different size which ranges from 16 to 256 points. In our experiments, the proposed architecture has the advantage of low power.