Multi-Modal Sensor Fusion using Hybrid Attention for Autonomous Driving

TL;DR

This paper introduces MMF-BEV, a hybrid radar-camera fusion framework using deformable attention for improved 3D object detection in autonomous driving, demonstrating superior performance over unimodal methods.

Contribution

The paper presents a novel radar-camera fusion architecture with deformable attention modules, along with a training strategy and sensor contribution analysis for interpretability.

Findings

MMF-BEV outperforms unimodal baselines in 3D detection accuracy.

Sensor contribution analysis reveals effective modality weighting at different distances.

The proposed method achieves competitive results against prior fusion approaches.

Abstract

Accurate 3D object detection for autonomous driving requires complementary sensors. Cameras provide dense semantics but unreliable depth, while millimeter-wave radar offers precise range and velocity measurements with sparse geometry. We propose MMF-BEV, a radar-camera BEV fusion framework that leverages deformable attention for cross-modal feature alignment on the View-of-Delft (VoD) 4D radar dataset [1]. MMF-BEV builds a BEVDepth [2] camera branch and a RadarBEVNet [3] radar branch, each enhanced with Deformable Self-Attention, and fuses them via a Deformable Cross-Attention module. We evaluate three configurations: camera-only, radar-only, and hybrid fusion. A sensor contribution analysis quantifies per-distance modality weighting, providing interpretable evidence of sensor complementarity. A two-stage training strategy - pre-training the camera branch with depth supervision, then jointly training radar and fusion modules stabilizes learning. Experiments on VoD show that MMF-BEV consistently outperforms unimodal baselines and achieves competitive results against prior fusion methods across all object classes in both the full annotated area and near-range Region of Interest.