Humanoid Education & Research Robot Manufacturer Software & Algorithm Provider Human-Machine Interaction
Flexion Develops Reinforcement Learning Platform to Address Teleoperation Issues in Robotics
Original from RoboticsBusinessReview.com: How to avoid the teleoperation trap in robotics development

Flexion Develops Reinforcement Learning Platform to Address Teleoperation Issues in Robotics

Flexion is developing a reinforcement learning and sim-to-real platform specifically for humanoid robots. Over the past 18 months, humanoid robotics companies have raised billions, primarily funding human operators to manage robots, which has led to a teleoperation and data challenge within the industry.

This reliance on teleoperation as a labor solution raises concerns about the long-term viability of training physical AI systems. The assumption that enough human demonstrations will enable robots to generalize across environments is questionable, especially given that teleoperation datasets are significantly smaller than those used for training language models, creating a growing data gap.

Looking ahead, the industry must address the limitations of teleoperation and the dependency on human input for robot training. If humanoid robots require continuous human demonstrations, the original vision of automation may be compromised. No further timeline was disclosed at the time of publication.

Editor's Note

The robotics industry is at a critical juncture as it grapples with the implications of teleoperation in training AI systems. The reliance on human operators not only raises questions about efficiency but also about the sustainability of such models in the face of evolving tasks and environments. As companies like Flexion innovate, the focus must shift towards developing robust training methodologies that reduce dependency on human input.

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

WAIC 2026: Embodied Intelligence Enters Industrial Deployment Era

WAIC 2026: Embodied Intelligence Enters Industrial Deployment Era

WAIC 2026 in Shanghai (July 17–20) saw 200+ embodied intelligence companies, 300+ live robots, and first confirmed humanoid deployments in logistics and pharmacy. RobotToday engineering analysis.

Market and Business News
Weekly Robotics Videos Showcase Advances in Humanoid Surgery and Quadrupedal Locomotion

Weekly Robotics Videos Showcase Advances in Humanoid Surgery and Quadrupedal Locomotion

This week's Video Friday highlights significant advancements in robotics, including a systematic evaluation of humanoid technology for laparoscopic surgical tasks. Researchers developed a humanoid-based teleoperation framework and assessed its performance through various studies, revealing both potential and challenges for clinical deployment. The importa...

Humanoid-robots Video-friday Quadruped-robots
Beijing Launches 2,000 Square Meter Hub for Humanoid Robot Training at Ice Ribbon

Beijing Launches 2,000 Square Meter Hub for Humanoid Robot Training at Ice Ribbon

On July 16, Beijing officially opened the Humanoid Robot Training Base at the Ice Ribbon, a 2,000 square meter innovation workshop. This facility will serve as a core venue for the upcoming second World Humanoid Robot Games in August. It was co-established by several organizations, including the Chaoyang Park Management Committee and Beijing Olympic Group...

Humanoid Robots Robot Training Innovation Workshop
Orbbec and Ant Group Unveil Advanced Data Collection Solutions at WAIC 2026

Orbbec and Ant Group Unveil Advanced Data Collection Solutions at WAIC 2026

At the 2026 World Artificial Intelligence Conference (WAIC) in Shanghai, Orbbec showcased its EGO RGB-D data collection platform in collaboration with Ant Group. This partnership aims to enhance data accuracy and stability for robotics applications by integrating self-developed depth chips and 3D vision hardware with spatial perception models. The signifi...

Data Collection Robotics 3D Vision
General Intuition Achieves $2.3 Billion Valuation with Innovative Robot Training Approach

General Intuition Achieves $2.3 Billion Valuation with Innovative Robot Training Approach

General Intuition, a New York-based company, has proposed a groundbreaking approach to training robots using millions of hours of gaming footage instead of vast amounts of real-world data. In June 2026, the company completed a $320 million Series A funding round, achieving a valuation of $2.3 billion, led by renowned investor Vinod Khosla. The significanc...

AI Robotics Gaming Technology
Ant Group Launches Open-AoE Framework for Embodied Intelligence Data Collection

Ant Group Launches Open-AoE Framework for Embodied Intelligence Data Collection

Ant Group, in collaboration with several universities and research institutions, has introduced the Open-AoE framework aimed at enhancing embodied intelligence data collection. This initiative addresses the scarcity of high-quality 3D operational data necessary for training robots, which often rely on limited and standardized datasets from controlled envi...

Embodied Intelligence Open Source Data Robot Training
Kinetix AI Introduces KAI Halo to Enhance Data Infrastructure for Robotics

Kinetix AI Introduces KAI Halo to Enhance Data Infrastructure for Robotics

As the robotics industry enters a phase of large-scale development, a critical question arises: how long does it take for newly collected real-world data to translate into actionable capabilities for robots? The data journey, from collection to deployment, is complex and any delays can hinder progress. Kinetix AI is addressing this challenge by connecting...

Embodied Intelligence Data Infrastructure Robotics
MIT Develops SceneSmith: AI System for Creating Realistic 3D Training Environments for Robots

MIT Develops SceneSmith: AI System for Creating Realistic 3D Training Environments for Robots

Researchers at MIT have developed SceneSmith, an AI-powered platform that generates realistic 3D indoor environments for robot training. This innovative system utilizes three collaborative AI agents to create detailed virtual spaces, enabling robots to practice everyday tasks safely and efficiently before real-world deployment. The significance of SceneSm...

AI and Robotics
MIT and Toyota Research Institute Unveil SceneSmith for Robot Household Training

MIT and Toyota Research Institute Unveil SceneSmith for Robot Household Training

MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) and the Toyota Research Institute have developed SceneSmith, an AI-powered system that allows robots to practice household tasks in a virtual environment. This system utilizes three visual language models to collaboratively create realistic 3D scenes, enabling robots to learn complex sk...

AI Robotics Virtual Reality
AI Agents Develop Virtual Environments for Essential Robot Training Data

AI Agents Develop Virtual Environments for Essential Robot Training Data

Robots are becoming more visible in public spaces, captivating onlookers. However, they still lack the versatility needed for tasks in kitchens or factories, primarily due to a significant data bottleneck. Similar to human learning, robots acquire skills through experience, but the process of physically training them in various environments is labor-inten...

Robotics
Manufacturers Must Prioritize Industrial Data Governance Before AI and Analytics

Manufacturers Must Prioritize Industrial Data Governance Before AI and Analytics

Manufacturers are generating unprecedented amounts of data through automation systems, capturing everything from process values to production metrics. However, many plants face challenges in quickly answering fundamental operational questions, such as equipment status during issues or alarm sequences. The root cause often lies in a lack of structured, con...

Factory / Analytics
MIT and Toyota Develop SceneSmith to Enhance Robot Training with AI-Generated Environments

MIT and Toyota Develop SceneSmith to Enhance Robot Training with AI-Generated Environments

MIT and the Toyota Research Institute have introduced SceneSmith, a system that utilizes AI agents to create realistic 3D environments for robot training. This innovation addresses the significant challenge of generating diverse simulation content, which is crucial for teaching robots various tasks in a cost-effective manner. The SceneSmith system employs...

Research Robotics Artificial intelligence

Related Suppliers

Dexteleop 灵御智能

Dexteleop Intelligence develops the TeleAvatar low-latency, high-precision teleoperation robot system. It features an omnidirectional chassis, 7-axis high-performance robotic arms, and force-controlled grippers, enabling skilled remote manipulation and presence for various applications.

IO-AI Tech 艾欧科技

IO Intelligence provides a one-stop solution for Embodied AI and robotics development. Their offerings include the TELEXPERIENCE III remote operation platform, human motion capture, and a data platform for large-scale data collection, annotation, and model training for dozens of robot types.

Robótica Industrial Proyectos México (MX)

Mexican firm with over 20 years of experience delivering industrial robot training courses for ABB, FANUC, KUKA, and YASKAWA platforms, from basic operation to advanced multi-move programming. Also provides 3D cell simulation, turnkey integration projects, and on-site instruction.

APT Manufacturing Solutions

APT Manufacturing Solutions is an exclusive FANUC Robot integrator for Robot Palletizing, Robot Welding and Robot Training equipment. We are a top level FANUC ASI (Authorized System Integrator) our Robot Welding equipment offers your choice of a Fronius or Miller power supply. We use both the FANUC COBOT or along with their traditional yellow robots

System Integrator

Averna GmbH

Averna provides automated test and quality assurance systems, including functional test equipment, measurement solutions, and data management platforms. Serves multi-sector industries with end-to-end test engineering for product validation, defect reduction, and lifecycle support.

Industrial Manufacturing Software & Algorithm Provider System Integrator

Deepen AI, Inc.

The product offers advanced data annotation and calibration solutions for autonomous systems, utilizing sensor fusion techniques, including LiDAR, camera, and radar integration. Key features include real-time productivity analytics, customizable data export, and built-in quality assurance workflows. The platform supports multi-sensor calibration, ensuring precise intrinsic and extrinsic parameter calculations, enabling efficient validation for safety-critical applications in robotics and AI.

Onward Robotics

FST Logistics has integrated Onward's Lumabot® autonomous mobile robots (AMRs) to enhance warehouse fulfillment efficiency. The system employs intelligent coordination algorithms for order prioritization, task assignment, and workflow optimization, facilitating seamless collaboration between human operators and robots. This technology aims to reduce operational downtime, improve pick efficiency, and maximize resource utilization without increasing labor costs.

AION Robotics Corporation

manufacturing floors of global industries. These robotics solutions leverage advanced sensors, AI algorithms, and IoT connectivity to enhance operational efficiency and safety. By integrating machine learning and data analytics, they provide predictive maintenance and real-time monitoring, optimizing workflows and reducing human error. This technology supports autonomous operations, enabling seamless interaction between machines and human operators in complex environments.

Quanser (Can)

Quanser is a Canadian company partnering with universities worldwide to deliver advanced laboratory ecosystems for control systems, robotics, and autonomous systems education, supporting physical AI and reinforcement learning research.

Education & Research Robot Manufacturer Control & Software Systems

Quilter

Quilter employs reinforcement learning algorithms to optimize PCB design workflows, significantly reducing layout cycles and validation times. The platform generates multiple candidate designs rapidly, leveraging physics-based simulations for accuracy. This approach enhances efficiency in the semiconductor industry, facilitating quicker iterations and enabling engineers to focus on core design tasks rather than non-essential layout activities.

Parallel Domain

Parallel Domain provides a synthetic data platform for training and testing autonomous systems. The API and SDK generate high-fidelity simulated sensor data — camera, LiDAR, radar — with full annotations, enabling perception teams to validate AI models at scale without real-world data collection.

OPAL-RT Technologies

Real-time simulation platforms (e.g., 4DV-SIM) for autonomous systems; models AV/robotics (perception, sensor fusion, control algorithms); supports ag off-road vehicles (e.g., LiDAR/camera simulation for manipulators/tasks); no hardware robots, focuses on HIL/MIL validation.

Education & Research Software & Algorithm Provider System Integrator