IEEE RA-L

UAV-MapFusion: RTK-Aligned Uncertainty-Aware Coarse-to-Fine Multi-Session UAV Mapping

UAV-MapFusion: RTK对齐的不确定性感知粗到细多会话无人机建图

Feng Pan, Chunran Zheng, Bing Xue, Yukang Cui, Jiayu Wen, Zhiyu Chen, Wei Wang

机构 * College of Automation, Harbin Engineering University（哈尔滨工程大学自动化学院） ； Department of Mechanical Engineering, University of Hong Kong（香港大学机械工程系） ； College of Mechatronics and Control Engineering, Shenzhen University（深圳大学机电与控制工程学院） 纠错

AI总结 提出一种不确定性感知的多会话点云地图合并与粗到细优化系统，通过RTK时空对齐和迭代平面因子优化，在抑制长距离漂移的同时保持局部几何精度。

Comments 8 pages, 5 figures, accepted by IEEE Robotics and Automation Letters (RA-L)

Scaling Nonlinear Optimization: Many Problems One GPU

扩展非线性优化：多问题单GPU

John Viljoen, Johanna Haffner, Masayoshi Tomizuka, Negar Mehr

机构 * Department of Mechanical Engineering, University of California, Berkeley（加州大学伯克利分校机械工程系） ； Department of Biosystems Science and Engineering, ETH Zürich（苏黎世联邦理工学院生物系统科学与工程系） 纠错

AI总结 提出首个基于JAX的GPU批处理NLP求解器jaxipm，通过异构迭代融合和迭代级批处理实现多问题并发求解，在四旋翼控制任务中吞吐量提升达32.85倍。

Comments 8 pages, 6 figures, ieeeconf style, submission for RA-L

CycleRL: Sim-to-Real Deep Reinforcement Learning for Robust Autonomous Bicycle Control

CycleRL：面向鲁棒自主自行车控制的仿真到现实深度强化学习

Gelu Liu, Teng Wang, Zhijie Wu, Junliang Wu, Songyuan Li, Xiangwei Zhu

机构 * School of Electronics and Communication Engineering and Shenzhen Key Laboratory of Navigation and Communication Integration, Sun Yat-sen University（电子工程学院和深圳导航通信集成重点实验室，中山大学） 纠错

AI总结 提出CycleRL框架，利用高保真仿真环境和域随机化，通过PPO算法实现自主自行车的平衡、速度跟踪和转向控制，在仿真中达到99.90%平衡成功率，并成功迁移到真实硬件。

Comments 8 pages, 7 figures, 8 tables. Accepted for publication in IEEE Robotics and Automation Letters (L-RA). See: https://ieeexplore.ieee.org/document/11568521

Journal ref IEEE Robotics and Automation Letters, vol. 11, no. 8, pp. 9343-9350, 2026

FC-Vision: Real-Time Visibility-Aware Replanning for Occlusion-Free Aerial Target Structure Scanning in Unknown Environments

FC-Vision：未知环境中无遮挡空中目标结构扫描的实时可见性感知重规划

Chen Feng, Yang Xu, Shaojie Shen

机构 * Dept. of Electronic and Computer Engineering, The Hong Kong University of Science and Technology（电子与计算机工程系，香港科学与技术大学） 纠错

AI总结 提出FC-Vision框架，通过双层分解（无遮挡视点修复和分段清洁感知连接）实时施加密集表面可见性约束，在保持覆盖率和效率的同时主动防止目标遮挡，实现55.32%的覆盖率提升和73.17%的遮挡率降低。

Comments Accepted by IEEE Robotics and Automation Letters. 8 pages, 8 figures, 3 tables. Code: https://github.com/FC-Family/FC-Vision. Video: https://www.youtube.com/watch?v=H3C42zlDOAI