SpINRv2: Implicit Neural Representation for Passband FMCW Radars

TL;DR

SpINRv2 is a neural framework that enhances high-fidelity 3D radar imaging by accurately modeling complex frequency responses and resolving ambiguities at high start frequencies, outperforming existing methods.

Contribution

It introduces a differentiable frequency-domain model with regularization techniques for improved volumetric reconstruction in FMCW radars, especially at high frequencies.

Findings

Outperforms classical and learning-based baselines

Achieves high spectral fidelity with reduced computational cost

Establishes a new benchmark for neural radar 3D imaging

Abstract

We present SpINRv2, a neural framework for high-fidelity volumetric reconstruction using Frequency-Modulated Continuous-Wave (FMCW) radar. Extending our prior work (SpINR), this version introduces enhancements that allow accurate learning under high start frequencies-where phase aliasing and sub-bin ambiguity become prominent. Our core contribution is a fully differentiable frequency-domain forward model that captures the complex radar response using closed-form synthesis, paired with an implicit neural representation (INR) for continuous volumetric scene modeling. Unlike time-domain baselines, SpINRv2 directly supervises the complex frequency spectrum, preserving spectral fidelity while drastically reducing computational overhead. Additionally, we introduce sparsity and smoothness regularization to disambiguate sub-bin ambiguities that arise at fine range resolutions. Experimental results show that SpINRv2 significantly outperforms both classical and learning-based baselines, especially under high-frequency regimes, establishing a new benchmark for neural radar-based 3D imaging.