On Hardware Implementations of Two-Dimensional Fast Fourier Transform for Radar Signal Processing

In the automotive industry, 77/79 GHz frequency- modulated continuous-wave (FMCW) radars have been vastly used for different driver-assistance and safety applications in modern vehicles. Range- Doppler processing relying on two- dimensional fast Fourier transform (2D- FFT) has emerged as one of the most efficient algorithms in extracting target range and velocity information in such systems. Having this in mind, this paper depicts several parameterized generators of radar signal 2D- FFT processing accelerators, implemented in Chisel hardware design language. Besides the basic range-Doppler algorithm, the proposed generators can provide slightly modified modules, tailored for range-angle calculations for obtaining information about target distance and azimuth. For cases with large amounts of intermediate data, the generated processors can utilize the ex-ternal SDRAM memory through an optional memory controller. The designs have been verified using both software simulations and hardware tests. Moreover, a demonstration system consisting of both range-Doppler and range-angle processing, implemented using a commercially-available radar front-end and an FPGA board, has been created for real-time radar imaging, thus proving the reliability and time-efficiency of the aforementioned designs.