Design and implementation of a fixed-point radix-4 FFT optimized for local positioning in wireless sensor networks

Local positioning is becoming an attractive feature for wireless sensor networks due to the variety of the existing industrial applications. The focus of developing autonomous wireless sensor nodes with localization functionality is with the localization accuracy and range, energy efficiency and the size of the sensor nodes. In this context special attention is paid to the sensor digital signal processing, where the main task is to perform a Fast Fourier Transform (FFT). This paper presents a design of the radix-4 FFT algorithm and its optimization with respect to hardware implementation for local positioning systems. In order to reduce the hardware requirements a fixed-point approach has been chosen. The algorithm was implemented, optimized and verified on a FPGA before starting the chip design. For the special case of a 4096-point FFT processing core the design has been realized as an ASIC using IBM 130nm CMOS technology.