We consider the task of underwater robot navigation for the purpose of collecting scientifically relevant video data for environmental monitoring. The majority of field robots that currently perform monitoring tasks in unstructured natural environments navigate via path-tracking a pre-specified sequence of waypoints. Although this navigation method is often necessary, it is limiting because the robot does not have a model of what the scientist deems to be relevant visual observations. Thus, the robot can neither visually search for particular types of objects, nor focus its attention on parts of the scene that might be more relevant than the pre-specified waypoints and viewpoints. In this paper we propose a method that enables informed visual navigation via a learned visual similarity operator that guides the robot's visual search towards parts of the scene that look like an exemplar image, which is given by the user as a high-level specification for data collection. We propose and evaluate a weakly supervised video representation learning method that outperforms ImageNet embeddings for similarity tasks in the underwater domain. We also demonstrate the deployment of this similarity operator during informed visual navigation in collaborative environmental monitoring scenarios, in large-scale field trials, where the robot and a human scientist collaboratively search for relevant visual content.
Cross-Modal Benchmarking for Robotic Perception in Natural Environments
CSIRO Robotics, CSIRO, Australia（CSIRO机器人研究所，CSIRO，澳大利亚）；University of Sydney (USyd), Australia（悉尼大学（USyd），澳大利亚）；Queensland University of Technology (QUT), Australia（昆士兰理工大学（QUT），澳大利亚）
ED3R: Energy-Aware Distributed Disaster Detection via Cooperative Agents in Robotic Systems
Dept. of Informatics and Telecommunications, National and Kapodistrian University of Athens（雅典大学信息学与电信系）；Huawei Heisenberg Research Center (Munich)（华为海森堡研究中心（慕尼黑））
Jet Propulsion Laboratory（喷气推进实验室）；California Institute of Technology（加州理工学院）；Swiss Federal Institute of Technology（瑞士联邦理工学院）；ETH Zürich（苏黎世联邦理工学院）；FieldAI Inc.（FieldAI公司）；Stanford University（斯坦福大学）；University of California Berkeley（加州大学伯克利分校）
TACO: Taming Diffusion for in-the-wild Video Amodal Completion
State Key Laboratory of General Artificial Intelligence, Peking University（北京大学通用人工智能国家重点实验室）；State Key Laboratory of General Artificial Intelligence, BIGAI（BIGAI 通用人工智能国家重点实验室）；Tsinghua University（清华大学）
MEBOW: Monocular Estimation of Body Orientation In the Wild
The Pennsylvania State University（宾夕法尼亚州立大学）；Amazon Lab126（亚马逊Lab126）
身体方向估计在许多应用中提供了关键的视觉线索，包括机器人和自动驾驶。当3-D姿态估计由于图像分辨率差、遮挡或不可区分的身体部位而难以推断时，这是特别期望的。我们提出了COCO-MEBOW（Monocular Estimation of Body Orientation in the Wild），这是一个新的大规模数据集，用于从单个野外图像中进行方向估计。COCO数据集的55 K图像中约130 K人体的身体方向标签已使用高效且高精度的注释管道收集。我们还验证了数据集的好处。首先，我们表明，我们的数据集可以大大提高性能和鲁棒性的人体方向估计模型，其发展是以前的规模和多样性的可用训练数据的限制。此外，我们提出了一种新的三维人体姿态估计的三源解决方案，其中3-D姿态标签，2-D姿态标签和我们的身体方向标签都用于联合训练。我们的模型在单目3-D人体姿势估计方面明显优于最先进的双源解决方案，其中训练仅使用3-D姿势标签和2-D姿势标签。这证实了MEBOW用于3-D人体姿态估计的重要优势，这是特别有吸引力的，因为身体方向的每个实例标记成本远远低于3-D姿态。这项工作展示了MEBOW在解决涉及理解人类行为的现实挑战方面的巨大潜力。关于这项工作的进一步信息可在https://chenyanwu.github.io/MEBOW/上查阅。
CineWild: Balancing Art and Robotics for Ethical Wildlife Documentary Filmmaking
Service Robotics Lab, Universidad Pablo de Olavide（帕布罗·德·奥拉维德大学服务机器人实验室）；Instituto de Investigación en Ingeniería de Aragón, Universidad de Zaragoza（阿拉贡工程研究所，萨拉戈萨大学）
Task-Informed Fidelity Management for Speeding Up Robotics Simulation
National Robotics Engineering Center（国家机器人工程中心）；Carnegie Mellon University（卡内基梅隆大学）；Jet Propulsion Laboratory（喷气推进实验室）；California Institute of Technology（加州理工学院）；Robotics Institute（机器人研究所）；Technion-Israel Institute of Technology（以色列理工学院）
Developing Flying Explorer for Autonomous Digital Modelling in Wild Unknowns
College of Civil Aviation, Nanjing University of Aeronautics and Astronautics（南京航空航天大学民航学院）；College of Engineering, South China Agriculture University（华南农业大学工程学院）
Decoding RobKiNet: Insights into Efficient Training of Robotic Kinematics Informed Neural Network
School of Mechanical Engineering, Shanghai Jiao Tong University（上海交通大学机械工程学院）；Intel Labs China（英特尔中国实验室）；Henan Academy of Sciences（河南省科学院）；Intel CCG FIS（英特尔中国区研究院）
School of Mechanical Engineering, Shanghai Jiao Tong University（上海交通大学机械与动力工程学院）；Intel Labs China（英特尔中国实验室）；Faculty of Mechanical and Electrical Engineering, Kunming University of Science and Technology（昆明理工大学机电工程学院）
Don't Fool Me Twice: Adapting to Adversity in the Wild with Experience-Driven Reasoning
Department of Engineering Design, Indian Institute of Technology, Madras（印度理工学院工程设计系，马德拉斯）；Robotics Institute, Carnegie Mellon University（卡内基梅隆大学机器人研究所）
Towards Robotic Companions: Understanding Handler-Guide Dog Interactions for Informed Guide Dog Robot Design
University of Massachusetts Amherst（马萨诸塞大学阿默斯特分校）；Indiana University Indianapolis（印第安纳大学印第安纳波利斯分校）；University of Maine（缅因大学）；University of Texas at Austin（德克萨斯大学奥斯汀分校）
