Spectral Domain Neural Reconstruction for Passband FMCW Radars

2025-06-09

AI Summary

To address voxel reconstruction distortion in high-start-frequency FMCW radar 3D imaging—caused by phase aliasing and sub-bin ambiguity—we propose the first fully differentiable frequency-domain forward model, embedding the radar’s physical response as a closed-form spectral representation into neural reconstruction. Our method integrates implicit neural representations (INRs), direct complex-spectrum supervision, and a sparse–total-variation joint regularizer that explicitly decouples sub-bin ambiguities. This design avoids time-domain inversion errors, explicitly models frequency-domain aliasing mechanisms, and preserves geometric consistency while maintaining full differentiability. Experiments demonstrate significant improvements over classical compressed sensing and state-of-the-art learning-based baselines at high frequencies, achieving new SOTA accuracy and robustness. Our work establishes an interpretable, optimization-friendly paradigm for neural radar 3D imaging.

Technology Category

• Computer Vision: Multi-modal Vision
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Feature Construction/Reformulation

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

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.

Problem

• High-fidelity volumetric reconstruction using FMCW radar
• Accurate learning under high start frequencies with phase aliasing
• Disambiguating sub-bin ambiguities at fine range resolutions

Innovation

• Differentiable frequency-domain forward model
• Implicit neural representation for scene modeling
• Sparsity and smoothness regularization