自动驾驶 - arXivDaily 专题

2607.28442 2026-07-31 cs.CV 新提交 70%

ViewMind3D: Modular View-Aware Inference for Training-Free 3D-QA

ViewMind3D：用于无需训练的3D问答的模块化视图感知推理

Ping-Kun Chiang, Kun-Ru Wu, Po-han Li, Sandeep Chinchali, Ufuk Topcu, Yu-Chee Tseng

专题命中感知：BEV(abstract,abstract_cn)；分类 cs.CV

AI总结该研究提出无需训练的模块化框架ViewMind3D，将3D-QA分解为四个组件，在ScanQA和SQA3D上实现竞争力性能，证明通用LLMs与VLMs的模块化编排可实现有效3D推理。

2607.28483 2026-07-31 cs.CV 新提交 57%

Towards Real-Time PixOOD: Efficient Anomaly Segmentation for Autonomous Vehicles

面向实时PixOOD：自动驾驶的高效异常分割

Luca de Martino, Federico Aromolo, Federico Nesti, Giorgio Buttazzo

机构 * Scuola Superiore Sant’Anna（圣安娜高等学校）

专题命中感知：autonomous driving(abstract)；分类 cs.CV

AI总结该研究针对自动驾驶和铁路领域，提出优化PixOOD的高效异常分割流水线，经TensorRT编译后在桌面GPU和嵌入式平台实现高帧率，解决了实时部署的计算成本问题。

Comments 12 pages, 2 figures, 3 tables. Accepted at the Efficient Deep Learning: Methods and Applications workshop, 35th International Conference on Artificial Neural Networks (ICANN 2026)

2607.28256 2026-07-31 cs.RO 新提交 57%

When Robots Exchange Meaning: A Demo of Goal-Oriented Semantic Communications for Collaborative Robotics

当机器人交换意义：面向协作机器人的目标导向语义通信演示

Peizheng Li, Xinyi Lin, Sajida Gufran, Adnan Aijaz

专题命中感知：LiDAR(abstract)；分类 cs.RO

AI总结本文提出一种机器人-边缘端语义通信测试平台，通过VQ-VAE等技术实现视觉压缩与语义建图，为协作机器人及相关6G网络研究提供了实用演示平台。

Comments 3 pages, 3 figures. This paper has been accepted for presentation as a demo paper at IEEE CSCN 2026

2604.19412 2026-07-31 cs.CV cs.CL 57%

VCE: A zero-cost hallucination mitigation method of LVLMs via visual contrastive editing

VCE：通过视觉对比编辑实现LVLMs的零成本幻觉抑制方法

Yanbin Huang, Yisen Li, Guiyao Tie, Xiaoye Qu, Pan Zhou, Hongfei Wang, Zhaofan Zou, Hao Sun, Xuelong Li

机构 * Huazhong University of Science and Technology（华中科技大学）； Institute of Artificial Intelligence (TeleAI)（人工智能研究院）； China Telecom（中国电信）

专题命中感知：autonomous driving(abstract)；分类 cs.CV

AI总结本文提出VCE方法，通过分析模型对对比视觉扰动的响应，利用SVD分解激活模式以抑制幻觉倾向，有效减少多基准测试中的物体幻觉，同时保持计算效率。

Comments ICASSP 2026

Journal ref 2026 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2026, pp. 16657-16661

2309.08324 2026-07-31 cs.RO 57%

Object-Oriented Grid Mapping in Dynamic Environments

Matti Pekkanen, Francesco Verdoja, Ville Kyrki

专题命中感知：occupancy(abstract)；分类 cs.RO

Comments IEEE-MFI 2024

Journal ref 2024 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI), Pilsen, Czech Republic, 2024, pp. 1-8

2607.27371 2026-07-31 eess.SY cs.SY 新提交 50%

Dynamics-matched Physical Reservoir Computing for Undersensed Traffic Prediction

动力学匹配物理储备池计算用于欠感知交通预测

Michael McCreesh, Rohit Gupta, Stephen L. Smith

专题命中感知：autonomous driving(abstract)

AI总结本研究提出将交通网络作为储备池的IIDM-RC方法，用于欠感知交通预测，在预测准确性和训练时间上优于ESNs与LSTM网络。

Comments 16 pages, 4 figures. Has been submitted for journal publication

1. 感知 6 篇