Deep learning (DL) has achieved great success in many applications, but it has been less well analyzed from the theoretical perspective. The unexplainable success of black-box DL models has raised questions among scientists and promoted the emergence of the field of explainable artificial intelligence (XAI). In robotics, it is particularly important to deploy DL algorithms in a predictable and stable manner as robots are active agents that need to interact safely with the physical world. This paper presents an analytic deep learning framework for fully connected neural networks, which can be applied for both regression problems and classification problems. Examples for regression and classification problems include online robot control and robot vision. We present two layer-wise learning algorithms such that the convergence of the learning systems can be analyzed. Firstly, an inverse layer-wise learning algorithm for multilayer networks with convergence analysis for each layer is presented to understand the problems of layer-wise deep learning. Secondly, a forward progressive learning algorithm where the deep networks are built progressively by using single hidden layer networks is developed to achieve better accuracy. It is shown that the progressive learning method can be used for fine-tuning of weights from convergence point of view. The effectiveness of the proposed framework is illustrated based on classical benchmark recognition tasks using the MNIST and CIFAR-10 datasets and the results show a good balance between performance and explainability. The proposed method is subsequently applied for online learning of robot kinematics and experimental results on kinematic control of UR5e robot with unknown model are presented.
EasyLabel: A Semi-Automatic Pixel-wise Object Annotation Tool for Creating Robotic RGB-D Datasets
TU Wien(维也纳技术大学);Aeolus Robotics, Inc.(风神机器人公司)
开发用于识别现实世界中的对象的机器人感知系统需要计算机视觉算法相对于预期的操作域进行仔细检查。这需要大量的地面实况数据来严格评估算法的性能。本文介绍了EasyLabel工具,用于在密集杂乱的场景中轻松获取像素级对象的高质量地面实况注释。在一个半自动的过程中,复杂的场景是渐进式构建的,EasyLabel利用深度变化在每一步提取精确的对象遮罩。我们使用这个工具来生成对象杂乱的室内数据集(OCID),该数据集捕获对象,背景,上下文,传感器到场景的距离,视角和照明条件的各种设置。OCID用于对现有的对象分割方法进行系统的比较。基线比较支持需要像素和对象明智的注释,以推进机器人视觉向现实的应用。这一见解揭示了EasyLabel和OCID在更好地理解机器人在现实世界中面临的挑战方面的有用性。版权所有20 XX IEEE。允许个人使用这些材料。在任何当前或未来的媒体中,必须获得IEEE的许可才能用于所有其他用途,包括出于广告或促销目的转载/重新发布此材料,创建新的集体作品,转售或重新分发到服务器或列表,或在其他作品中重新使用本作品的任何受版权保护的组件。
Istituto Italiano di Tecnologia (IIT)(意大利技术研究院(IIT));Institut de Robòtica i Informàtica Industrial, CSIC-UPC(西班牙国家研究委员会-加泰罗尼亚理工大学机器人与信息自动化研究所);Heriot-Watt University(赫瑞瓦特大学)
Time-Unified Diffusion Policy with Action Discrimination for Robotic Manipulation
National Key Laboratory of Human-Machine Hybrid Augmented Intelligence(人机混合增强智能国家重点实验室);National Engineering Research Center for Visual Information and Applications(视觉信息与应用国家工程研究中心);Institute of Artificial Intelligence and Robotics, Xi'an Jiaotong University(西安交通大学人工智能与机器人研究院);School of Computing and Artificial Intelligence, Southwestern University of Finance and Economics(西南财经大学计算机与人工智能学院)
Real-Time Instrument Segmentation in Robotic Surgery using Auxiliary Supervised Deep Adversarial Learning
NUS Graduate School for Integrative Sciences and Engineering (NGS)(新加坡国立大学综合科学与工程研究生院(NGS));National University of Singapore(新加坡国立大学);University of Moratuwa(莫拉图瓦大学);NIT Trichy(印度国立理工学院蒂鲁吉拉伯利分校)
StreamingWAM: Streaming World Action Modeling via Chunk-Wise Staircase Flow Matching
University of Chinese Academy of Sciences(中国科学院大学);Beijing Zhongguancun Academy(北京中关村学院);Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);University of Science and Technology of China(中国科学技术大学);Nankai University(南开大学);Harbin Institute of Technology(哈尔滨工业大学);Eastern Institute of Technology, Ningbo(宁波工程学院)
针对现有WAM机器人控制方法未同时解决推理效率与动作连续性的问题,提出StreamingWAM流扩散框架,在多机器人任务上提升动作连续性,降低重规划延迟并减少对执行 horizon K 的敏感性。
DigiForest: Digital Analytics and Robotics for Sustainable Forestry
Dipartimento di Ingegneria Industriale, University of Trento(工业工程系,特伦托大学);Faculty of Engineering, Free University of Bozen-Bolzano(工程学院,博兹纳-博尔扎诺自由大学);Faculty of Agricultural, Environmental and Food Sciences, Free University of Bozen-Bolzano(农业、环境与食品科学学院,博兹纳-博尔扎诺自由大学);Competence Centre for Mountain Innovation Ecosystems(山地创新生态系统研究中心);School of Informatics, University of Edinburgh(信息学院,爱丁堡大学);Technical University of Munich, School of Computation, Information and Technology(慕尼黑技术大学,计算、信息与技术学院);Department of Engineering Cybernetics, Norwegian University of Science and Technology(工程 cybernetics 系,挪威科学技术大学);Center for Robotics, University of Bonn(机器人中心,波恩大学);Swiss Federal Research Institute WSL(瑞士联邦研究 institute WSL);Oxford Robotics Institute, Department of Engineering Science, University of Oxford(牛津大学机器人研究所,工程科学系);Mobile Robotics Lab, ETH Zürich, Department of Mechanical and Process Engineering(移动机器人实验室,苏黎世联邦理工学院机械与过程工程系)
An Efficient Numerical Function Optimization Framework for Constrained Nonlinear Robotic Problems
Automation and Control Institute (ACIN), TU Wien(自动化与控制研究所(ACIN));Institute of Robotics and Mechatronics, German Aerospace Center (DLR)(机器人与机电研究所,德国航空航天中心(DLR))
Learning 6-DoF Fine-grained Grasp Detection Based on Part Affordance Grounding
Center for X-Mechanics, School of Aeronautics and Astronautics, Zhejiang University(浙江大学航空航天学院X力学中心);Information Hub, Data Science and Analytics Thrust, The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州)信息中心、数据科学与分析方向);Advanced Manufacturing Lab, Huawei Technologies(华为技术先进制造实验室);Research Institute, UBTECH Robotics Inc.(UBTECH机器人研究院);School of Information and School of Smart Governance, Renmin University of China(中国人民大学信息学院和智能治理学院);School of Engineering, Westlake University and the Institute of Advanced Technology, Westlake Institute for Advanced Study(西湖大学工程学院和西湖先进科技研究院);Department of Electrical and Computer Engineering, Dalhousie University(达尔豪斯大学电气与计算机工程系)
Shaping Expressiveness in Robotics: The Role of Design Tools in Crafting Embodied Robot Movements
CHArt Laboratory, Department of Psychology, University Paris8(巴黎第八大学心理学系CHArt实验室);Lab of INIT Robots, Department of Mechanical Engineering, École de technologie supérieure(École de technologie supérieure机械工程系INIT机器人实验室);Department of Dance, Université du Québec à Montréal(魁北克大学蒙特利尔分校舞蹈系);Costech Laboratory, Department of Humanities and Social Sciences, Université technologique de Compiègne(Compiègne技术大学人文与社会科学系Costech实验室)
Neural Minimum-Distance Estimation for Collision-Aware Operation of Multi-Arm Laparoscopy Surgical Robots Through Learning-from-Simulation
Surgical Performance Enhancement and Robotics (SuPER) Centre, Department of Surgery(外科性能增强与机器人中心(SuPER)中心,外科部);The Wilfred and Joyce Posluns Centre for Image Guided Innovation & Therapeutic Intervention (PCIGITI)(威廉与乔伊斯·波斯伦中心(PCIGITI)影像引导创新与治疗干预中心);The Hospital for Sick Children (SickKids)(儿童医院(SickKids))