NRI: Collaborative Research: Experiential Learning for Robots: From Physics to Actions to Tasks
NRI:协作研究:机器人的体验式学习:从物理到动作再到任务
基本信息
- 批准号:1637949
- 负责人:
- 金额:$ 64.8万
- 依托单位:
- 依托单位国家:美国
- 项目类别:Standard Grant
- 财政年份:2016
- 资助国家:美国
- 起止时间:2016-10-01 至 2020-08-31
- 项目状态:已结题
- 来源:
- 关键词:
项目摘要
Recent advances in machine learning coupled with unprecedented archives of labeled data are advancing machine perception at a remarkable rate. However, applying these advances to robotics has not advanced as quickly because learning for robotics requires both active interaction with the physical world, and the ability to generalize over a variety of task contexts. This project addresses this knowledge gap through the development of new learning methods to produce experience-based models of physics. In this approach, an object or category specific model of physics is learned directly from perceptual data rather than deploying general-purpose physical simulation methods. These physical models will support both direct control of action - for example pouring a liquid into a container, and the learning of the physical effects of sequences of actions - for example planning to handle fluids in a laboratory. More generally, these methods will provide a means for robots to learn how to handle fluids, soft materials, and other complex physical phenomena. The proposed experiential learning framework will build on recent advances in deep neural networks. The key problem is to learn the mappings between raw perceptual and control data via a low-dimensional implicit physics space representing a perception-based physical model of how an object acts in the environment. Three directions will be investigated: 1) the development of experiential physics models for object interaction and fluid flow that have strong predictive capabilities, 2) creating mappings directly from experiential models to control of actions such as pouring or moving an object, 3) the assembly of local experience-based controllers into complex tasks from interactive demonstration. Additionally, the project will develop unique data sets that include physical models, simulations, data components, and learned components that other groups can access and build on to enable comparative research similar to what has emerged in machine perception.
机器学习的最新进展加上前所未有的标记数据档案,正在以惊人的速度推进机器感知。然而,将这些进步应用于机器人技术并没有那么快,因为机器人技术的学习需要与物理世界进行积极的互动,以及在各种任务环境中进行概括的能力。该项目通过开发新的学习方法来解决这一知识差距,以产生基于经验的物理模型。在这种方法中,对象或类别特定的物理模型直接从感知数据中学习,而不是部署通用的物理模拟方法。这些物理模型将支持对动作的直接控制-例如将液体倒入容器中,以及对动作序列的物理效果的学习-例如计划在实验室中处理流体。更一般地说,这些方法将为机器人提供一种学习如何处理流体、软材料和其他复杂物理现象的方法。 拟议的体验式学习框架将建立在深度神经网络的最新进展之上。关键问题是通过低维隐式物理空间来学习原始感知数据和控制数据之间的映射,该物理空间表示对象在环境中如何行为的基于感知的物理模型。将研究三个方向:1)开发具有强大预测能力的对象交互和流体流动的经验物理模型,2)直接从经验模型创建映射到诸如倾倒或移动对象等动作的控制,3)将基于本地经验的控制器组装到交互式演示的复杂任务中。此外,该项目还将开发独特的数据集,包括物理模型、模拟、数据组件和学习组件,其他团队可以访问和构建这些数据集,以实现类似于机器感知中出现的比较研究。
项目成果
期刊论文数量(13)
专著数量(0)
科研奖励数量(0)
会议论文数量(0)
专利数量(0)
Using Data-Driven Domain Randomization to Transfer Robust Control Policies to Mobile Robots
- DOI:10.1109/icra.2019.8794343
- 发表时间:2019-05
- 期刊:
- 影响因子:0
- 作者:Matthew Sheckells;Gowtham Garimella;Subhransu Mishra;Marin Kobilarov
- 通讯作者:Matthew Sheckells;Gowtham Garimella;Subhransu Mishra;Marin Kobilarov
Evaluating Methods for End-User Creation of Robot Task Plans
- DOI:10.1109/iros.2018.8594127
- 发表时间:2018-10
- 期刊:
- 影响因子:0
- 作者:Chris Paxton;Felix Jonathan;Andrew Hundt;Bilge Mutlu;Gregory Hager
- 通讯作者:Chris Paxton;Felix Jonathan;Andrew Hundt;Bilge Mutlu;Gregory Hager
Visual Robot Task Planning
- DOI:10.1109/icra.2019.8793736
- 发表时间:2018-03
- 期刊:
- 影响因子:0
- 作者:Chris Paxton;Yotam Barnoy;Kapil D. Katyal;R. Arora;Gregory Hager
- 通讯作者:Chris Paxton;Yotam Barnoy;Kapil D. Katyal;R. Arora;Gregory Hager
The CoSTAR Block Stacking Dataset: Learning with Workspace Constraints
- DOI:10.1109/iros40897.2019.8967784
- 发表时间:2018-10
- 期刊:
- 影响因子:0
- 作者:Andrew Hundt;Varun Jain;Chia-Hung Lin;Chris Paxton;Gregory Hager
- 通讯作者:Andrew Hundt;Varun Jain;Chia-Hung Lin;Chris Paxton;Gregory Hager
Do what i want, not what i did: Imitation of skills by planning sequences of actions
做我想做的事,而不是我做过的事:通过规划行动序列来模仿技能
- DOI:10.1109/iros.2016.7759556
- 发表时间:2016
- 期刊:
- 影响因子:0
- 作者:Paxton, Chris;Jonathan, Felix;Kobilarov, Marin;Hager, Gregory D.
- 通讯作者:Hager, Gregory D.
{{
item.title }}
{{ item.translation_title }}
- DOI:
{{ item.doi }} - 发表时间:
{{ item.publish_year }} - 期刊:
- 影响因子:{{ item.factor }}
- 作者:
{{ item.authors }} - 通讯作者:
{{ item.author }}
数据更新时间:{{ journalArticles.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ monograph.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ sciAawards.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ conferencePapers.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ patent.updateTime }}
Gregory Hager其他文献
Task re-encoding in vision-based control systems
基于视觉的控制系统中的任务重新编码
- DOI:
10.1109/cdc.1997.650586 - 发表时间:
1997 - 期刊:
- 影响因子:0
- 作者:
Wen;J. Hespanha;A. Morse;Gregory Hager - 通讯作者:
Gregory Hager
Active Multispectral Illumination and Image Fusion for Retinal Microsurgery
用于视网膜显微手术的主动多光谱照明和图像融合
- DOI:
10.1007/978-3-642-13711-2_2 - 发表时间:
2010 - 期刊:
- 影响因子:0
- 作者:
R. Sznitman;Seth D. Billings;D. Rother;D. Mirota;Yi Yang;J. Handa;P. Gehlbach;Jin U. Kang;Gregory Hager;R. Taylor - 通讯作者:
R. Taylor
Deep Hierarchical Multi-label Classification of Chest X-ray Images
胸部 X 射线图像的深度分层多标签分类
- DOI:
- 发表时间:
2018 - 期刊:
- 影响因子:0
- 作者:
Haomin Chen;S. Miao;Daguang Xu;Gregory Hager;Adam P. Harrison - 通讯作者:
Adam P. Harrison
CoSTAR in Surgery : A Cross-platform User Interface for Surgical Robot Task Specification
手术中的 CoSTAR:用于手术机器人任务规范的跨平台用户界面
- DOI:
- 发表时间:
2017 - 期刊:
- 影响因子:0
- 作者:
Baichuan Jiang;Chris Paxton;P. Kazanzides;Gregory Hager - 通讯作者:
Gregory Hager
What Tasks can be Performed with an Uncalibrated Stereo Vision System?
使用未校准的立体视觉系统可以执行哪些任务?
- DOI:
10.1023/a:1008111128520 - 发表时间:
1999 - 期刊:
- 影响因子:19.5
- 作者:
J. Hespanha;Z. Dodds;Gregory Hager;A. Morse - 通讯作者:
A. Morse
Gregory Hager的其他文献
{{
item.title }}
{{ item.translation_title }}
- DOI:
{{ item.doi }} - 发表时间:
{{ item.publish_year }} - 期刊:
- 影响因子:{{ item.factor }}
- 作者:
{{ item.authors }} - 通讯作者:
{{ item.author }}
{{ truncateString('Gregory Hager', 18)}}的其他基金
RI: Medium: Collaborative Research: Towards Practical Encoderless Robotics Through Vision-Based Training and Adaptation
RI:中:协作研究:通过基于视觉的训练和适应实现实用的无编码机器人技术
- 批准号:
1900952 - 财政年份:2019
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
Planning Grant: Engineering Research Center for Augmentation Systems and Intelligent Support Technologies for Aging (ASISTa-ERC)
规划资助:衰老增强系统与智能支持技术工程研究中心(ASISTa-ERC)
- 批准号:
1840446 - 财政年份:2018
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
RI: Medium: Robots That Learn From Description Through Synthesis and Analysis
RI:媒介:通过综合和分析从描述中学习的机器人
- 批准号:
1763705 - 财政年份:2018
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
Doctoral Consortium at the 18th International Symposium on Robotics Research
第18届国际机器人研究研讨会博士联谊会
- 批准号:
1749288 - 财政年份:2017
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
NRI-Large: Collaborative Research: Multilateral Manipulation by Human-Robot Collaborative Systems
NRI-Large:协作研究:人机协作系统的多边操纵
- 批准号:
1227277 - 财政年份:2012
- 资助金额:
$ 64.8万 - 项目类别:
Continuing Grant
The Computing Community Consortium II
计算社区联盟 II
- 批准号:
1136993 - 财政年份:2012
- 资助金额:
$ 64.8万 - 项目类别:
Cooperative Agreement
International: A US-Germany Research Collaboration on Systems for Computer-Integrated Healthcare
国际:美德计算机集成医疗保健系统研究合作
- 批准号:
1065092 - 财政年份:2011
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
The Third Computing Innovation Fellows Project
第三届计算创新研究员项目
- 批准号:
1136996 - 财政年份:2011
- 资助金额:
$ 64.8万 - 项目类别:
Continuing Grant
CPS:Medium:Hybrid Systems for Modeling and Teaching the Language of Surgery
CPS:中:手术语言建模和教学的混合系统
- 批准号:
0931805 - 财政年份:2009
- 资助金额:
$ 64.8万 - 项目类别:
Continuing Grant
相似海外基金
NRI/Collaborative Research: Robotic Disassembly of High-Precision Electronic Devices
NRI/合作研究:高精度电子设备的机器人拆卸
- 批准号:
2422640 - 财政年份:2024
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
NRI/Collaborative Research: Robust Design and Reliable Autonomy for Transforming Modular Hybrid Rigid-Soft Robots
NRI/合作研究:用于改造模块化混合刚软机器人的稳健设计和可靠自主性
- 批准号:
2327702 - 财政年份:2023
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
Collaborative Research: NRI: Understanding Underlying Risks and Sociotechnical Challenges of Powered Wearable Exoskeleton to Construction Workers
合作研究:NRI:了解建筑工人动力可穿戴外骨骼的潜在风险和社会技术挑战
- 批准号:
2410255 - 财政年份:2023
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
NRI: FND: Collaborative Research: DeepSoRo: High-dimensional Proprioceptive and Tactile Sensing and Modeling for Soft Grippers
NRI:FND:合作研究:DeepSoRo:软抓手的高维本体感受和触觉感知与建模
- 批准号:
2348839 - 财政年份:2023
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
Collaborative Research: NRI: Reducing Falling Risk in Robot-Assisted Retail Environments
合作研究:NRI:降低机器人辅助零售环境中的跌倒风险
- 批准号:
2132936 - 财政年份:2022
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
NRI/Collaborative Research: Robust Design and Reliable Autonomy for Transforming Modular Hybrid Rigid-Soft Robots
NRI/合作研究:用于改造模块化混合刚软机器人的稳健设计和可靠自主性
- 批准号:
2133019 - 财政年份:2022
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
Collaborative Research: NRI: Remotely Operated Reconfigurable Walker Robots for Eldercare
合作研究:NRI:用于老年护理的远程操作可重构步行机器人
- 批准号:
2133075 - 财政年份:2022
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
Collaborative Research: NRI: Smart Skins for Robotic Prosthetic Hand
合作研究:NRI:机器人假手智能皮肤
- 批准号:
2221479 - 财政年份:2022
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
Collaborative Research: NRI: Integration of Autonomous UAS in Wildland Fire Management
合作研究:NRI:自主无人机在荒地火灾管理中的整合
- 批准号:
2132798 - 财政年份:2022
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant
NRI/Collaborative Research: Robot-Assisted Feeding: Towards Efficient, Safe, and Personalized Caregiving Robots
NRI/合作研究:机器人辅助喂养:迈向高效、安全和个性化的护理机器人
- 批准号:
2132847 - 财政年份:2022
- 资助金额:
$ 64.8万 - 项目类别:
Standard Grant