Enhancing Fourier-based Doppler Resolution with Diffusion Models
This addresses a domain-specific radar processing challenge, offering an incremental improvement over traditional FFT methods.
The paper tackles the problem of insufficient Doppler resolution in radar systems for detecting slow-moving targets by using diffusion models to enhance range-Doppler maps, demonstrating effective separation of closely spaced targets.
In radar systems, high resolution in the Doppler dimension is important for detecting slow-moving targets as it allows for more distinct separation between these targets and clutter, or stationary objects. However, achieving sufficient resolution is constrained by hardware capabilities and physical factors, leading to the development of processing techniques to enhance the resolution after acquisition. In this work, we leverage artificial intelligence to increase the Doppler resolution in range-Doppler maps. Based on a zero-padded FFT, a refinement via the generative neural networks of diffusion models is achieved. We demonstrate that our method overcomes the limitations of traditional FFT, generating data where closely spaced targets are effectively separated.