Can BEV Perception Gracefully Degrade under Sensor Failures?

2026-05-29

AI Summary

This work addresses the vulnerability of multimodal bird’s-eye-view (BEV) perception systems to sensor failures, which often cause severe performance degradation due to static fusion mechanisms. To mitigate this issue, the authors propose Grace-BEV, a novel framework that introduces an active modality reliability assessment mechanism for the first time. Specifically, a TrustGate Router dynamically quantifies the trustworthiness of each modality, and a FailSafe Fusion Block adaptively adjusts feature fusion in BEV space based on these reliability estimates. Combined with a three-stage training strategy and modality dropout, Grace-BEV achieves lightweight, plug-and-play graceful degradation. Experiments on nuScenes-R/C demonstrate substantial robustness gains: when LiDAR completely fails, mAP recovers from 0.0% to 34.7%, while also improving mAP by 1.4% on clean data.

Abstract

Despite the remarkable success of multi-modal bird's-eye view (BEV) perception in autonomous driving, current systems exhibit a critical vulnerability: existing fusion mechanisms are highly brittle to sensor corruptions, often causing catastrophic performance degradation. This vulnerability largely stems from the fact that standard fusion frameworks typically integrate multi-modal representations in a static manner, leading to a precipitous performance collapse under missing or corrupted modalities. In contrast, we show that graceful degradation is achievable through active modality reliability assessment. To this end, we present Grace-BEV, a lightweight and plug-and-play framework that enforces active reliability awareness during multi-modal fusion. Instead of relying on computationally expensive cross-modal interactions, Grace-BEV leverages the aligned BEV space to explicitly assess modality trustworthiness via a TrustGate Router and dynamically recalibrate feature integration using the FailSafe Fusion Block. Furthermore, we devise a Three-Phase Training strategy with Modality Dropout to prevent modality dominance and encourage balanced cross-modal learning under unreliable inputs. Extensive experiments on nuScenes-R and nuScenes-C show that Grace-BEV maintains robust performance across diverse corruption settings. Notably, under catastrophic LiDAR failures where standard baselines collapse to 0.0% mean Average Precision (mAP), Grace-BEV restores performance to as high as 34.7% mAP. Moreover, it improves clean accuracy by up to 1.4%, achieving a strong trade-off between robustness and efficiency.

Problem

• BEV perception
• sensor failures
• multi-modal fusion
• graceful degradation
• modality corruption

Innovation

• graceful degradation
• multi-modal fusion
• BEV perception
• modality reliability
• sensor failure robustness