ACCOR: Attention-Enhanced Complex-Valued Contrastive Learning for Occluded Object Classification Using mmWave Radar IQ Signals

TL;DR

Results demonstrate the benefits of integrating complex-valued deep learning, attention, and contrastive learning for mmWave radar-based occluded object classification.

Abstract

Millimeter-wave (mmWave) radar provides robust sensing under adverse conditions and can penetrate thin materials for non-visual perception in industrial and robotic settings. Recent work with MIMO mmWave radar has demonstrated its ability to penetrate cardboard packaging for occluded object classification. However, existing models leave room for extensions and improvements across different sensing frequencies. Building on recent work with MIMO radar for occluded object classification, we propose ACCOR, an attention-enhanced complex-valued contrastive learning approach for radar, enabling robust occluded object classification. ACCOR processes complex-valued IQ radar signals via a complex-valued CNN backbone, a multi-head attention layer and a hybrid loss. The hybrid loss combines a weighted cross-entropy term with a supervised contrastive term. We extend an existing 64 GHz dataset with a new 67 GHz subset and evaluate performance across both bands. ACCOR achieves 96.60 % accuracy at 64 GHz and 93.59 % at 67 GHz on 10 objects, surpassing prior radar-specific and adapted image models. Results demonstrate the benefits of integrating complex-valued deep learning, attention, and contrastive learning for mmWave radar-based occluded object classification.

Figures (8)

• Figure 1: Virtual channels formed by each antenna pair in the L-shaped $20 \times 20$ antenna array.
• Figure 2: Sensing Setup: a top-mounted mmWave radar sensing the contents of a closed packaging box.
• Figure 3: Image of the measurement setup from stefan.
• Figure 4: ACCOR model consisting of FFT preprocessing, a CNN backbone, and a multi-head attention layer, trained with the proposed hybrid loss function.
• Figure 5: Complex-valued CNN backbone with three layers, kernel size 5, and average pooling, kernel size 2.