Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces Variance-Guided Spatial Attention Fusion (VG-SAF), a novel framework designed to enhance the robustness of end-to-end driving systems under asymmetric sensor degradation. By integrating a physically grounded augmentor for simulating sensor failures, modality-specific experts for reliability estimation, and a hybrid attention mechanism, VG-SAF effectively suppresses unreliable features and improves decision-making in challenging conditions. The approach significantly outperforms existing methods on the CARLA Longest6 benchmark, demonstrating enhanced driving performance across various degradation scenarios.
VG-SAF boosts driving robustness by effectively filtering out unreliable sensor data, achieving superior performance even under severe degradation conditions.
End-to-end multimodal driving has progressed rapidly by fusing camera and LiDAR streams. Existing pipelines remain fragile under asymmetric sensor degradation, where either an entire modality or only a localized region is corrupted while other regions remain useful. The key difficulty is not simply to add an uncertainty head, but to obtain dense reliability supervision, calibrate this reliability against physical fault severity, and use it before unreliable features bias the planner. We propose Variance-Guided Spatial Attention Fusion (VG-SAF), in which dense heteroscedastic reliability estimates act as interpretable spatial gates. The framework couples three components. First, a physically grounded augmentor simulates representative camera and LiDAR failures and emits a continuous spatial mask, providing dense supervision without additional annotation. Second, modality-specific experts predict per-pixel reliability scales through cross-branch dense distillation in log space, enforcing a monotone severity-to-scale response. Third, calibrated reliability maps drive a hybrid attention mechanism that suppresses unreliable cells with a local spatial gate and arbitrates between modalities through a cross-modal trust softmax. A Laplace uncertainty head emits a systemic waypoint uncertainty scale that signals severe or combined sensor degradation, including severities outside the training ranges. On the CARLA Longest6 benchmark, VG-SAF consistently improves closed-loop robustness over the baselines across camera-only, LiDAR-only, and joint degradation regimes, as measured by driving score, route completion, and infraction score.