11-19 更新 实时
首页/ 算力/ 正文

AI 智能体在虚拟世界中测试并改进机器人代码

算力 来源:TechXplore 2026-11-09 22:11 0 阅读 机器翻译
AI 智能体在虚拟世界中测试并改进机器人代码

由 Sadie Harley 编辑,Robert Egan 审阅

本文已按照 Science X 的编辑流程与政策进行审核。编辑在确保内容可信度的同时,重点标注了以下属性:

添加到首选来源

Goldberg 及其团队发布了一套行业标准的基准测试,用于评估机器人在各类工业任务中的表现,图中展示了其中的示例。图片来源:Ken Goldberg

近几个月来,能够进行推理、规划并使用外部工具完成多步骤任务的人工智能智能体,已经重塑了计算机编程、在线购物、餐厅预订和网络攻击等领域。

加州大学伯克利分校机器人学家、工程学教授 Ken Goldberg 表示,新模型使 AI 智能体团队能够开发机器人控制程序、在模拟环境中测试这些程序,并在出现问题时修改代码。

Goldberg、NVIDIA 机器人工程总监 Spencer Huang 以及一个研究团队将这种新方法称为“智能体式机器人”。他们表示,这或许能绕过该领域的两大障碍:训练机器人所需的庞大数据量,以及为其编程所需的大量工程投入——而这两者都需要耗费巨量的实体与技术资源才能完成。

“智能体式机器人是机器人控制方式的一次根本性转变,”Goldberg 说。“团队无需再收集海量数据并逐一手工调校每一项行为,而是可以由智能体团队快速生成、测试并优化机器人程序,让人类工程师能够理解这些程序的工作方式,以便进行微调和验证。”

Goldberg最近入选美国国家工程院,他将在机器人学习大会(Conference on Robot Learning)上展示这一新方法以及数月测试的结果。他和团队还将发布开源代码,供其他机器人研究者开展实验,该代码基于NVIDIA面向开发者的Isaac开放仿真框架构建;同时他们还发布了一套行业标准基准测试,用于评估机器人在各类工业任务中的表现。这项研究已发布在arXiv预印本服务器上。

Goldberg和Huang表示,目标是让机器人编程更快、更具适应性,同时保留构建可靠工业系统所需的结构。

“机器人技术正在快速进步,但为真实环境开发机器人仍然进展缓慢、数据密集且难以规模化,”Huang说。“与UC Berkeley合作,NVIDIA正在帮助推进一种新方法,将智能体式 AI、仿真以及面向工业任务的开放基准测试结合起来,帮助开发者在部署前快速测试并改进机器人行为。”

绕开机器人领域数据缺口的一条路径

传统上,工程师可以直接编写软件,告诉机器人如何感知物体、规划运动并控制自身动作,这一过程被称为基于模型的机器人技术。或者,采用无模型的机器人技术,工程师可以通过模仿学习,训练机器人从演示中学习行为。第一种方法需要大量的工程投入,第二种则需要海量数据和时间。

问题在于,目前没有简单的方法来收集如此海量的信息用于训练机器人。为了重置公众预期,Goldberg 去年曾表示,“10 万年的数据差距”将使机器人无法像 AI 智能体在过去几年里迅速获得语言流畅度那样,快速掌握现实世界的技能。毕竟,那些 AI 智能体是在海量互联网文本和图像数据上训练出来的,而机器人领域并没有与之相当的数据。

“需要收集的数据量之大令人望而生畏,我为此感到沮丧,”Goldberg 说。“智能体式机器人技术填补了基于模型的方法与无模型方法之间缺失的一环。这是一条无需任何额外数据的前进之路。”

在仿真中构建并优化机器人技能

今年早些时候,他和团队开始尝试改变机器人代码的组织方式。他们采用一个名为 Graph-as-Policy 的框架,把机器人任务拆分成一项项独立技能,比如寻找物体、规划抓取动作,以及在不发生碰撞的情况下移动夹爪。Goldberg 将其比作一块块积木,可以拼装成更大得多的东西。

随后,AI 智能体生成这些技能,并把它们连接成一个机器人程序。借助 NVIDIA 的 Isaac 仿真库,这些智能体在多种任务变体中自主演练该程序,找出失败之处,并在部署前改进方案。

Goldberg 说,OpenAI 最近发布的 Astra 模型让这一过程更加前景可期。他的博士后学生 Eric Kaiyuan Chen 和 Shuangyu Xie 让 Astra 接入了他们的仿真界面,并提供了一段真实机器人的短视频,任务是用海绵擦掉金属杆上的墨水。不到一小时,这个 AI 工具就创建并微调出一个物理仿真,精准复现了这段视频,包括海绵非常逼真的形变。

随后,研究人员把该仿真和他们的 Graph-as-Policy 框架一并交给 AI。AI 生成了一套机器人控制程序,并在仿真中反复测试和改进。当研究人员把最终代码运行在真实机器人上时,它奏效了。

Goldberg 说,在一次试验中,机器人推倒了金属杆,因为它无法可靠地推断其质量。这次失误为智能体提供了可供复盘的新证据,也凸显了智能体式机器人学的优势:AI 可以利用失败尝试的视频来改进其仿真、修改机器人的控制程序,并测试新的方案。

“就是这一刻,我起了鸡皮疙瘩,”Goldberg 说。

仓储场景已有收益,家庭助手仍遥遥无期

总部位于伯克利的 Ambi Robotics 专注于物流与仓储订单履行,今年早些时候应用了这些新方法,用于分析包裹分拣数据并测试机器人行为的更新。测试显示,每台机器人每年可分拣的包裹量增加了超过 15,000 件;这一改动目前已部署到该公司在美国 30% 的机器人车队上。

“AI 驱动的工业机器人必须持续提升可靠性和性能,”Ambi Robotics 联合创始人兼首席技术官 Jeff Mahler 表示。“Agentic Robotics 将生产数据转化为机器人行为更新,提升了我们在美国全部机器人机队的吞吐量,同时让工程师始终掌握控制权。”

Goldberg 表示,在任务重复且规模化的工业场景中,这一技术的意义已经初现前景。他说,别指望家用助手很快就能叠衣服——至少目前还不行——但让任务中的关键环节具备自我改进能力,已是该领域的一大飞跃。

对他和他的学生来说,这也是一个激动人心的时刻。他说,在他 40 年的机器人研究生涯中,只有两个时刻让他产生过类似感受:1993 年万维网(World Wide Web)面向公众发布,以及 2022 年 ChatGPT 首次亮相。

“智能体式机器人技术将经典控制理论与物理学,与快速演进的智能体式编程结合在一起,”Goldberg 说。“它速度快,对实际工程应用而言很实用。在我看来,这是一次真正的范式转变。”

Kaiyuan Chen 等,GaP: A Graph-as-Policy Multi-Agent Self-Learning Harness For Variational Automation Tasks,arXiv(2026)。DOI: 10.48550/arxiv.2607.05369

期刊信息:arXiv

这篇报道的作者是谁?

生命科学与生态学学士。拥有微生物学实验室背景,并具备石油、天然气和可再生能源行业的医药新闻经验。完整简介 →

数学生物学学士,创意写作硕士。游历广泛,对科学与语言有独到见解。完整简介 →

引文:AI 智能体在部署前于虚拟世界中测试并改进机器人代码(2026,11月9日),2026年11月9日检索自 https://techxplore.com/news/2026-11-ai-agents-robot-code-virtual.html

本文受版权保护。除为私人学习或研究目的进行的合理使用外,未经书面许可,不得复制本文任何部分。内容仅供参考。

获取即时摘要文本(Gist)

AI 智能体在物理仿真中生成了模块化机器人控制程序,并对其进行测试和改进,随后成功迁移到真实机器人上。在仓库测试中,行为更新使每台机器人的年分拣能力提升超过 15,000 件包裹,并已部署至该公司美国 30% 的机器人机队中。

本摘要由 LLM 自动生成。完整免责声明

译文由英文原文机器翻译,术语按行业惯例统一,可在文末展开原文核对。

查看英文原文(AI agents test and improve robot code in virtual worlds before deployment)

edited by Sadie Harley, reviewed by Robert Egan

This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility:

Add to Preferred Sources

Recent advances in AI are creating a 'paradigm shift' for robotics Goldberg and his team released an industry-standard benchmark for evaluating the performance of robots across industrial tasks, examples of which are shown here. Credit: Ken Goldberg

In recent months, artificial intelligence agents that can reason, plan and use external tools to accomplish multistep tasks have already reshaped computer coding, online shopping,dinner reservations and cyberattacks.

Ken Goldberg, a UC Berkeley roboticist and professor of engineering, said new models make it possible for teams of AI agents to develop robot-control programs, test them in simulated environments and revise the code when something doesn't work.

Goldberg, Spencer Huang, director of robotics engineering at NVIDIA, and a team of researchers call the new approach "agentic robotics." They said it could offer a way around two obstacles in the field: the enormous amounts of data needed to train robots and the intense engineering efforts required to program them, both of which can demand vast amounts of physical and technical resources to accomplish.

"Agentic robotics is a fundamental shift in how robots are controlled," Goldberg said. "Instead of collecting massive amounts of data and hand-tuning every behavior, teams of agents can quickly generate, test and refine robot programs that allow human engineers to interpret how the programs work for fine-tuning and verification."

Goldberg, who was recently inducted into the National Academy of Engineering, is presenting the new approach and results from months of testing at the Conference on Robot Learning. He and his team are also releasing open-source code for other roboticists to experiment with, built on NVIDIA's Isaac open simulation framework for developers, as well as an industry-standard benchmark for evaluating the performance of robots across industrial tasks. The work is published on the arXiv preprint server.

Goldberg and Huang said the goal is to make robot programming faster and more adaptable while preserving the structure needed for reliable industrial systems.

"Robots are advancing quickly, but developing them for real-world environments remains slow, data-intensive and difficult to scale," Huang said. "Working with UC Berkeley, NVIDIA is helping advance a new approach combining agentic AI, simulation and open industrial task-focused benchmarks to help developers rapidly test and improve robot behavior before deployment."

A route around robotics' data gap

Traditionally, engineers can directly program the software that tells robots how to perceive objects, plan movements and control their actions, a process called model-based robotics. Or, with model-free robotics, they can train robots to learn behaviors from demonstrations through imitation learning. The first approach requires extensive engineering, while the second needs enormous amounts of data and time.

The problem is there's no easy way to collect that much information to train robots on. In an attempt to reset public expectations last year, Goldberg said that a "100,000-year data gap" would prevent robots from gaining real-world skills as quickly as AI agents have gained language fluency over the past several years. Those AI agents, after all, were trained on an immense amount of internet text and image data for which there is no parallel in robotics.

"I was discouraged by the daunting amount of data that we would need to collect," Goldberg said. "Agentic robotics provides a missing link between model-based and model-free methods. It's a way forward that doesn't require any additional data."

Building and refining robot skills in simulation

Earlier this year, he and his team began experimenting with a change to the way the robot code is organized. Using a framework they call Graph-as-Policy, robot tasks are segmented into individual skills, such as finding an object, planning a grasp and moving the gripper without collisions. Goldberg likened it to individual building blocks that fit together into something much larger.

Then, AI agents generate and connect these skills into a robot program. Using NVIDIA's Isaac simulation libraries, the agents autonomously rehearse the program across task variations, identify where failures happen and refine the approach before deployment.

OpenAI's recent release of its Astra model made the process even more promising, Goldberg said. His postdoctoral students Eric Kaiyuan Chen and Shuangyu Xie gave Astra access to their simulation interface and a short video of a real robot whose assignment was using a sponge to wipe ink off a metal bar. In less than an hour, the AI tool created and fine-tuned a physics simulation that accurately reproduced the video, including very realistic deformations of the sponge.

The researchers then gave the AI that simulation and their Graph-as-Policy framework. The AI generated a robot-control program and repeatedly tested and improved it in simulation. When the researchers ran the resulting code on a real robot, it worked.

In one trial, Goldberg said, the robot pushed over the metal bar because it did not reliably infer its mass. The misstep created new evidence for the agent to review and illustrated the edge provided by agentic robotics: The AI could use video of the failed attempt to improve its simulation, revise the robot's control program and test a new approach.

"This is when I got goosebumps," Goldberg said.

Warehouse gains, with home assistants still distant

Berkeley-based Ambi Robotics, which specializes in logistics and warehouse order fulfillment, applied the new methods earlier this year to analyze parcel-sorting data and test updates to robot behavior. Tests showed an increased capacity of over 15,000 packages sorted annually per robot; the change is now deployed across 30% of the company's U.S. fleet.

"AI-powered industrial robots must continuously improve reliability and performance," said Jeff Mahler, Ambi Robotics' co-founder and chief technology officer. "Agentic Robotics converted production data into robot behavior updates, increasing throughput across our U.S. fleet while keeping engineers in control."

Goldberg said the implications are already promising in industrial settings, where tasks are repetitive and at scale. He said not to expect home assistants to fold laundry—yet—but the ability to self-improve key elements of tasks is a major leap forward in the field.

It's also an exciting time for him and his students. He said he can think of only two moments in his 40-year robotics career that have elicited similar feelings: the public release of the World Wide Web in 1993 and the debut of ChatGPT in 2022.

"Agentic robotics combines classical control theory and physics with rapidly evolving agentic coding," Goldberg said. "It's fast and practical for real engineering applications. My sense is that it's a true paradigm shift."

Kaiyuan Chen et al., GaP: A Graph-as-Policy Multi-Agent Self-Learning Harness For Variational Automation Tasks, arXiv (2026). DOI: 10.48550/arxiv.2607.05369

Journal information:arXiv

Who's behind this story?

Sadie Harley

BSc Life Sciences & Ecology. Microbiology lab background with pharmaceutical news experience in oil, gas, and renewable industries. Full profile →

Bachelor's in mathematical biology, Master's in creative writing. Well-traveled with unique perspectives on science and language. Full profile →

Citation: AI agents test and improve robot code in virtual worlds before deployment (2026, November 9) retrieved 9 November 2026 from https://techxplore.com/news/2026-11-ai-agents-robot-code-virtual.html

This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.

Get Instant Summarized Text (Gist)

AI agents generated, tested and refined modular robot-control programs in physics simulations, with successful transfer to a real robot. In warehouse tests, behavior updates increased annual sorting capacity by over 15,000 packages per robot and were deployed across 30% of a company's U.S. fleet.

This summary was automatically generated using LLM. Full disclaimer

This site uses cookies to assist with navigation, analyse your use of our services, collect data for ads personalisation and provide content from third parties. By using our site, you acknowledge that you have read and understand our Privacy Policy and Terms of Use.

原文为英文,正文由自动抓取并保留原文。

本文由采集器自动同步

原文发布于 TechXplore(AI agents test and improve robot code in virtual worlds before deployment),标题、摘要与正文由机器翻译,术语按行业惯例统一,未经人工逐句校订。

阅读原文 ↗