4DR360: State Reasoning for Joint 3D Detection and Occupancy Prediction with 4D Radar-Camera Fusion

Summary

4DR360 is a 4D radar-camera framework for 360-degree full-scene perception in autonomous driving, proposed by Xiaokai Bai, Lianqing Zheng, Runwei Guan, Songkai Wang, Siyuan Cao, and Hui-liang Shen (arXiv:2607.09629). The method addresses the sparsity of 4D millimeter-wave radar returns by fusing them with camera data, and models semantic occupancy as a persistent scene state rather than a terminal output. The framework follows a cross-modal state reasoning paradigm with two key modules: State-guided BEV Enhancement (SBE), which strengthens intra-frame BEV representations, and Doppler-guided Temporal Fusion (DTF), which preserves state evidence over longer time horizons. The authors also extend the ManTruckScenes dataset with a unified cross-dataset detection-occupancy protocol. This work targets reliable autonomous driving perception that jointly captures foreground objects and dense semantic layout.