Education & Research Software & Algorithm Provider Human-Machine Interaction Perception & Vision
Chen Pu of Yuan Ke Vision Discusses Transitioning Robot Training Data from 2D to 4D with High Precision
Original from leaderobot.com: Yuan Ke Vision's Chen Pu: Advancing Robot Training Data from 2D to 4D with Sub-Millimeter Precision

Chen Pu of Yuan Ke Vision Discusses Transitioning Robot Training Data from 2D to 4D with High Precision

On September 23, Chen Pu, Vice President of Product Development at Yuan Ke Vision, highlighted the importance of high-quality real data for robot training during a seminar in Beijing. He emphasized the need to elevate data from 2D to 4D and improve precision from centimeter to sub-millimeter levels to build a robust 4D data foundation for robotics.

This transition is crucial as the industry faces challenges such as weak model generalization and inadequate scene adaptability. Chen noted that current training methods often rely on limited 2D video data, which fails to capture the complexities of three-dimensional space and temporal changes, hindering robots' ability to understand and interact with their environments effectively.

Looking ahead, Yuan Ke Vision aims to address these challenges by developing new data collection and training paradigms that enhance dimensionality, enrich modalities, and improve precision. Chen pointed out that while video data serves as a foundation, it is often too simplistic, and the lack of tactile data remains a significant gap in the industry. No further timeline was disclosed at the time of publication.

Editor's Note

The robotics industry is increasingly focusing on the quality of training data as a critical factor for the advancement of embodied intelligence. The shift from 2D to 4D data collection is essential for enabling robots to perform complex tasks in real-world environments. Companies must prioritize the development of high-precision data acquisition technologies to remain competitive in this evolving landscape.

RobotToday Initiative

Robotics needs a service framework.

RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.

Share

Related Articles/News

TachinGlove Snap Revolutionizes Haptic Data Collection with Quick-Detach Feature

TachinGlove Snap Revolutionizes Haptic Data Collection with Quick-Detach Feature

Tujian Technology has introduced the TachinGlove Snap, a flexible electronic skin haptic data collection glove featuring a quick-detach module. This innovation allows users to easily swap the data collection module without interrupting the workflow, enhancing efficiency in high-frequency testing and multi-user scenarios. The development of haptic data col...

Haptic Technology Data Collection Wearable Devices
Three Companies Shift Focus Away from Collecting Real Machine Data

Three Companies Shift Focus Away from Collecting Real Machine Data

Three companies have decided to stop collecting real machine data, indicating a significant shift in their operational strategies. This change reflects a broader trend in the industry, where organizations are exploring alternative methods for data utilization and management. The implications of this decision could reshape how machine data is perceived and...

Robotics Automation AI
Light Origins Unveils Light-O1 Foundation Model to Enhance Robot Learning Capabilities

Light Origins Unveils Light-O1 Foundation Model to Enhance Robot Learning Capabilities

Light Origins has launched Light-O1, its first general-purpose embodied foundation model, which utilizes human actions extracted from internet videos to provide a reusable framework for robots to learn various tasks. The company trained six versions of the 4-billion-parameter model on multimodal tokens, achieving significant reductions in prediction error...

AI AI Funding & Investment AI Infrastructure & Compute
RobotToday Weekly September 21 – 25, 2026

RobotToday Weekly September 21 – 25, 2026

Cognex buys RealSense, Qualcomm acquires PickNik, the IFR counts 5 million factory robots, AGIBOT deploys 300 robots at Chimelong, and Tekever raises $580 million.

RobotToday Weekly
Mifengpai Unveils Data Crowdsourcing Initiative for Robot Training Through Daily Activities

Mifengpai Unveils Data Crowdsourcing Initiative for Robot Training Through Daily Activities

On September 23, Mifengpai launched a global initiative to crowdsource data for robot training, showcasing over 50,000 real environments and 5,000 tasks. The event highlighted the need for extensive data on everyday actions, which are crucial for training robots but have not been systematically recorded. Mifengpai's approach combines hardware, an app, and...

Data Crowdsourcing Robot Training AI Technology
Research Reveals Manufacturers Struggle with Data Connectivity Across Systems

Research Reveals Manufacturers Struggle with Data Connectivity Across Systems

Recent research indicates that fewer than half of manufacturers have fully integrated data across their design, production, quality, and business systems. Despite 77% of manufacturers claiming their operations are connected, only 40% have achieved full data connectivity, as reported in Hexagon’s 2026 America's State of Manufacturing Report. This disconnec...

Factory / Digital Transformation
Brett Adcock Predicts Humanoids Will Require More Data and Compute Than LLMs

Brett Adcock Predicts Humanoids Will Require More Data and Compute Than LLMs

Figure CEO Brett Adcock has stated that scaling humanoid robots will necessitate hundreds of billions of dollars in investment and more data and computational power than large language models (LLMs). In a recent interview, he outlined a four-stage development process for humanoids, emphasizing the need for capable hardware and advanced AI control architec...

Figure AI manufacturing
Brain-Computer Interface Funding Surpasses 7 Billion Yuan in Six Months, Non-Invasive Solutions Lead Market

Brain-Computer Interface Funding Surpasses 7 Billion Yuan in Six Months, Non-Invasive Solutions Lead Market

In the first half of 2026, funding for brain-computer interface (BCI) technology exceeded 7 billion yuan, with over 60 financing events reported. Non-invasive solutions dominate the market, accounting for 86% of the total, while invasive and semi-invasive solutions make up the remaining 14%. The Chinese government has recognized BCI as a key investment ar...

Brain-Computer Interfaces Healthcare Technology Non-Invasive Solutions
23 Key Scenarios for Embodied Intelligence Valued at 250 Billion Yuan Annually

23 Key Scenarios for Embodied Intelligence Valued at 250 Billion Yuan Annually

The State Council of China has issued guidelines to accelerate the cultivation and application of high-value scenarios for embodied intelligence, highlighting the country's vast market and complete industrial system. This initiative aims to foster strategic and economically significant scenarios that meet the needs of technology, industry, and consumption...

Embodied Intelligence High-Value Scenarios Technological Innovation
Five One Vision's Embodied Intelligence Business Achieves Over 200 Million Yuan in Orders

Five One Vision's Embodied Intelligence Business Achieves Over 200 Million Yuan in Orders

On September 17, Five One Vision's stock opened nearly 3% higher, reaching 56.50 HKD, following the announcement of significant order data. Since the launch of its embodied intelligence business, APER, on August 18, the company has accumulated orders exceeding 200 million Yuan in less than a month. The customer base includes companies in embodied data col...

Embodied Intelligence Data Collection Robotics
Innovation Works Venture Expo 2026: 58 Pittsburgh Startups Pitch to Investors

Innovation Works Venture Expo 2026: 58 Pittsburgh Startups Pitch to Investors

58 health, deep tech and robotics startups pitched at Innovation Works Venture Expo 2026 in Pittsburgh. Full company list, program and 2025 funding data.

Market and Business News
Shangpin Home Introduces Open-source WorldSimReady-Home Dataset for Robotics Training

Shangpin Home Introduces Open-source WorldSimReady-Home Dataset for Robotics Training

Shangpin Home, in collaboration with Tangyuan Technology, has launched the WorldSimReady-Home simulation dataset aimed at addressing the challenges of robotic training in complex home environments. This open-source dataset includes 100,000 square meters of high-fidelity home scenes, 10,000 interactive assets, and 1,000 standardized robotic simulation task...

Robotics Training Simulation Data Home Automation

Related Suppliers

Maniformer 觅觅科技

AgiBot-incubated Chinese platform supplying embodied-AI training data, real and simulation datasets, and the MEgo data-collection hardware line.

Synthium

Synthium is a Singapore-based, Antler-backed startup that crowdsources gamified simulations to generate training data for humanoid robot AI models.

CalTa Inc.

Tokyo digital twin startup (est. 2021, JR East JV) offering TRANCITY Nebula cloud platform converting drone/robot video into 4D infrastructure data.

SiLC Technologies, Inc.

Monrovia, California silicon-photonics company founded by Dr. Mehdi Asghari; its Eyeonic FMCW LiDAR line reaches sub-millimeter depth precision.

VinDynamics

Vietnamese robotics company under VinGroup developing robot motors, robotic hands, and physical AI data infrastructure from TechnoPark Tower, Hanoi.

reNAND.ai

Full-stack AI data infrastructure platform for autonomous driving, robotics, and generative AI.

Hubei Optics Valley Dongzhi Embodied Intelligence Technology Co., Ltd. 湖北光谷东智具身智能技术有限公司

Wuhan, China embodied-intelligence firm founded 2025, backed by HEC Group, Zhiyuan Robotics, and Peking University; makes Lumen humanoid and service robots.

Sichuan Embodied Intelligent Robot Technology Co., Ltd 四川具身人形机器人科技有限公司

Chengdu humanoid-robot startup building the Tianxingzhe series and the emotionally interactive Aiqiu-AIQ companion robot.

NOKOV Metric Technology 度量科技

Beijing optical motion capture system maker; sub-millimeter precision 3D tracking for robotics, UAVs, VR, biomechanics, and embodied AI.

NINOX 360

Redwood City, California computer vision firm and A3 member since 2022; makes the Surface Inspector 3D scanner for corrosion measurement.

Industrial Manufacturing Logistics & Supply Chain Robot Manufacturer

Tachin Technology Co., Ltd. 途见科技(北京)有限公司

Beijing flexible-electronics startup founded 2020; makes tactile-sensing electronic skin, gloves and smart-cockpit touch systems for robots and vehicles.

Sycsense Technology 赛感科技(深圳)有限公司

Shenzhen maker of flexible capacitive tactile sensors and robot electronic skin for humanoid robots, industrial manipulators, and medical wearables.

Robot Manufacturer Manipulation & Grasping Human-Machine Interaction

inJoin the RobotToday community on LinkedIn

Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.