Variance-Guided Spatial Attention Fusion for Robust End-to-End Driving under Asymmetric Sensor Degradation
面向非对称传感器退化场景下鲁棒端到端驾驶的方差引导空间注意力融合
机构 * The Hong Kong Polytechnic University(香港理工大学) ; Anhui University(安徽大学) ; Shenzhen Research Institute of Northwestern Polytechnical University(西北工业大学深圳研究院)
专题命中 端到端驾驶 :end-to-end driving(title);LiDAR(abstract,abstract_cn);分类 cs.RO、cs.CV
AI总结 提出VG-SAF框架,通过物理增强器、模态专家与混合注意力机制,在CARLA Longest6基准上提升非对称传感器退化下的端到端驾驶鲁棒性。
Comments 17 pages, 9 figures, and 4 tables, including supplementary material. Submitted to IEEE Transactions on Vehicular Technology