This paper presents an active search trajectory synthesis technique for autonomous mobile robots with nonlinear measurements and dynamics. The presented approach uses the ergodicity of a planned trajectory with respect to an expected information density map to close the loop during search. The ergodic control algorithm does not rely on discretization of the search or action spaces, and is well posed for coverage with respect to the expected information density whether the information is diffuse or localized, thus trading off between exploration and exploitation in a single objective function. As a demonstration, we use a robotic electrolocation platform to estimate location and size parameters describing static targets in an underwater environment. Our results demonstrate that the ergodic exploration of distributed information (EEDI) algorithm outperforms commonly used information-oriented controllers, particularly when distractions are present.
Autonomous Robotic Tissue Palpation and Abnormalities Characterisation via Ergodic Exploration
Department of Information Engineering and Computer Science, University of Trento(信息工程与计算机科学系,特伦托大学);Department of Industrial Engineering, University of Trento(工业工程系,特伦托大学)
Balanced Collaborative Exploration via Distributed Topological Graph Voronoi Partition
College of Electrical Engineering, Zhejiang University(浙江大学电气工程学院);Jinhua Institute of Zhejiang University(浙江大学金华研究院);National Key Laboratory of Human-Machine Hybrid Augmented Intelligence, Xi’an Jiaotong University(西安交通大学人机混合增强智能国家级重点实验室);National Key Laboratory of Industrial Control Technology, Zhejiang University(浙江大学工业控制技术国家级重点实验室)
Distributed multi-robot potential-field-based exploration with submap-based mapping and noise-augmented strategy
Singapore University of Technology and Design(新加坡科技设计大学);Inria Grenoble, Grenoble Alpes University(法国国家信息与自动化研究所格勒诺布尔分部,格勒诺布尔阿尔卑斯大学);School of Electrical and Electronic Engineering, Nanyang Technological University(南洋理工大学电气与电子工程学院)
Learning Object-conditioned Exploration using Distributed Soft Actor Critic
Robotics at Google(谷歌机器人部门)
对象导航被定义为在复杂的、未探索的环境中导航到给定标签的对象。在其一般形式中,这个问题给机器人技术带来了几个挑战:在搜索对象和低级控制时对未知环境进行语义探索。在这项工作中,我们研究了对象引导的探索和低级别的控制,并提出了一个端到端的训练导航策略,实现了0.68的成功率和0.58的SPL对看不见的,视觉上复杂的扫描真实的家园。我们提出了一种高度可扩展的非策略强化学习算法的实现,分布式Soft Actor Critic,它允许系统在8个GPU上在24小时内利用98M经验步骤。我们的系统学习控制差分驱动移动基地在模拟中,从机器人平台上常用的高维观测堆栈。学习的策略能够从现实环境中的纯经验中学习到对象引导的探索性行为和低级控制。
A Distributed Multi-UGV Exploration Framework With Loop-Aware Planning and Descriptor-Aided Localization in Resource-Limited Environments
School of Mechanical Engineering, Beijing Institute of Technology(北京理工大学机械与车辆学院);China North Artificial Intelligence & Innovation Research Institute, Collective Intelligence & Collaboration Laboratory (CIC)(中国北方人工智能与创新研究院集体智能与协作实验室);Zhengzhou Intelligent Technology Research Institute, Beijing Institute of Technology(北京理工大学郑州智能科技研究院)
Rapidly Converging Time-Discounted Ergodicity on Graphs for Active Inspection of Confined Spaces
Department of Mechanical Engineering, University of Washington(机械工程系,华盛顿大学);Department of Electrical & Computer Engineering, University of Washington(电气与计算机工程系,华盛顿大学);Naval Undersea Warfare Center(海军水下作战中心);Department of Industrial & Systems Engineering and the Department of Mechanical Engineering, University of Washington(工业与系统工程系和机械工程系,华盛顿大学)
针对连续空间遍历轨迹难跟踪的问题,提出图上时间折扣遍历性度量与凸规划马尔可夫链,并结合 SLAM 贝叶斯检验用于受限空间异常检测。
Distributed Multi-Robot Multi-Target Simultaneous Search and Tracking in an Unknown Non-convex Environment
School of Electrical and Automation Engineering, Nanjing Normal University(电气与自动化工程学院,南京师范大学);Department of Mechanical Engineering, Temple University(机械工程系, Temple大学)
Behaviorally Heterogeneous Multi-Agent Exploration Using Distributed Task Allocation
Contextual Robotics Institute and the Department of Mechanical and Aerospace Engineering, UC San Diego(情境机器人研究所和机械与航空航天工程系,加州大学圣地亚哥分校);U.S. Army Combat Capabilities Development Command, Army Research Laboratory(美国陆军战斗能力发展司令部,陆军研究实验室)