RoboSignal Daily · 2026-10-31 · Saturday
ROBOSIGNAL
Kinematic-Tactile Aligned Dexterous Retargeting with Contact Wrench Space Alignment
This research presents a three-stage framework for dexterous retargeting that aligns human and robot hand kinematics and tactile information through contact wrench space. The method achieves 96.7% replay success in real robot experiments, demonstrating effective transfer of physical interaction data for manipulation tasks.
Robot Models
LeRobot Releases Cosmos3-Nano Policy Conversion
LeRobot has released a policy conversion of the NVIDIA Cosmos3-Nano base checkpoint, providing model components for robot policy post-training. The conversion does not include task-specific fine-tuning and requires a target dataset to define frame rate and camera size.
Research
arXiv Robotics — research abstractsKinematic-Tactile Aligned Dexterous Retargeting with Contact Wrench Space Alignment
This research presents a three-stage framework for dexterous retargeting that aligns human and robot hand kinematics and tactile information through contact wrench space. The method achieves 96.7% replay success in real robot experiments, demonstrating effective transfer of physical interaction data for manipulation tasks.
Tech Xplore — RoboticsArmored caterpillars offer robot designers lessons in crossing slippery glass
Researchers from the University of Naples Federico II and the University of Palermo studied how caterpillars move across various surfaces, including glass, and developed a mathematical model to simulate their locomotion. By adding artificial armor and weights, they improved the caterpillars' ability to crawl on smooth surfaces, offering insights for the design of bio-inspired soft robots.
arXiv Robotics — research abstractsELEVATE: Agentic Controller Acquisition from Task Failure for Quadrupedal Manipulation
ELEVATE is a framework that uses persistent task failures to identify and acquire missing whole-body control capabilities for quadrupedal manipulators. Starting from a single pretrained controller, it autonomously learns controllers for wall-supported elevation, supported manipulation, and wall descent across seven simulation tasks. These controllers achieve 98.4%-100% success rates in tasks beyond the initial controller's workspace and transfer to a physical robot with 18/20 successes in rack retrieval and 16/20 in floor-to-rack placement.
2026 Heavy-Duty Collaborative Robots Surge: Evolution from Palletizing to Embodied Intelligence
In 2026, the collaborative robot market experienced significant changes, with 30kg+ load products beginning to enter mainstream enterprise product lines in large numbers. The article analyzes the application demands of heavy-duty collaborative robots in scenarios such as palletizing and food handling, explores technical trends including load-to-weight ratio optimization, high-speed motion control, and safety perception systems, and points out that the industry is evolving from light-load to higher load capacities and stronger adaptability to demanding working conditions.
Zhiyuan Releases Embodied Model GO-3 Preview, Tops Three Industry Rankings
On October 29, 2026, Zhiyuan released the GO-3 Preview, an embodied model based on the native embodied architecture WRAM, achieving unified task reasoning, future prediction, and action generation. GO-3 Preview ranked first in three major rankings: RoboDojo, RoboTwin, and RoboColiseum, and plans to fully open-source by the end of 2026. The model continuously iterates through pre-training, post-training, and deployment, enabling end-to-end scalable learning and enhancing robots' adaptability and execution capabilities in complex tasks.
From Video to Action: PowerTech Breaks Through Single-View 3D Reconstruction Accuracy, Turning Human Experience into Robot Training Data
PowerTech has developed a method to convert single-view RGB videos into high-precision 3D action data for robotic training. The approach involves high-accuracy 3D hand pose reconstruction, autonomous truth systems for validation, and automated pipelines for large-scale processing. This enables the transformation of human demonstrations into structured data that can be used to train robots, with applications in imitation learning and embodied perception. The system achieves sub-millimeter accuracy in hand pose reconstruction and supports real-time mapping to robotic arms and dexterous hands.
AI Robotics May Need 10 Million Hours of Interaction Data to Reach Production-Level Capabilities
A Goldman Sachs report highlights that general-purpose AI robotics may require approximately 10 million hours of high-quality, multimodal interaction data to achieve production-level capabilities. The report identifies key challenges in physical interaction data, spatial perception, and real-time execution, suggesting that advances in AI models alone may not be sufficient for commercial viability without improvements in physical precision and reliability.
We walk without worrying about what’s behind us. Why should humanoids?
The ETH Robotic Systems Lab introduces 'Anticipated Perception for Humanoid Locomotion,' a new method that prioritizes terrain perception for the feet's upcoming landing points, aiming to enhance stability and navigation in humanoids.
Robots & Hardware
Tech Xplore — RoboticsRobot Moth Grows UV Sensor Using Artificial Circulatory System
Researchers from the University of Tartu and the Italian Institute of Technology developed a moth-shaped robot that uses an artificial circulatory system to deliver chemical building blocks and grow a light sensor where UV light strikes. The robot's body is 3D-printed with microchannels that allow fluid to flow, enabling the synthesis of a conductive polymer sensor. This approach allows the robot to create new hardware on-site, enhancing its adaptability to unpredictable environments.
IEEE Spectrum — RoboticsVideo Friday: Spot Demonstrates Whole-Body Manipulation
IEEE Spectrum's Video Friday highlights Spot's whole-body manipulation capabilities, along with other robotics projects such as humanoid mobility, robotic hands, and specialized industrial robots. The video also includes demonstrations of AI-driven drone racing and energy-efficient amphibious robots, showcasing the breadth of current robotics innovation.
Industry
Robotics & Automation NewsBleu Robotics Raises €5 Million to Scale Humanoid Robot Deployment in Factories
Bleu Robotics, a Paris-based startup, has raised €5 million in a pre-seed round to scale its software solution for industrial humanoid robots. The company aims to reduce the full deployment process to under an hour by using abductive learning, which enables robots to learn new tasks from as few as ten operator demonstrations. This approach eliminates the need for large datasets and long training cycles, allowing for cost-effective and scalable deployment in manufacturing settings.
Briefs
- Who Is the Agent Talking To? Participant-Aware Role Switching for Embodied Student-Support DialoguearXiv Robotics — research abstracts
- ScheduleGen: Demonstration Generation via Skill Scheduling for Multi-Arm Imitation LearningarXiv Robotics — research abstracts
- Embodied at Sea: Imitation Learning for Multi-Task Underwater ManipulationarXiv Robotics — research abstracts
- Xinyan Robotics Launches Embodied Family Agent 'Bubbo 1'Robotics — Chinese web discovery
- Jungheinrich and Automha Partner on Warehouse AutomationRobotics & Automation News
- Aitai Demonstrates 3C Embodied Production Line, Shifting from Showcasing Parameters to Demonstrating RhythmHigh-tech Robotics — WeChat
- Introducing I3-3L: Iterative Interactive Imitation Learning for Whole-Body Mobile ManipulationFei-Fei Li — X Robotics
- From HIL to RIL. Vehicle Grade Technology Enters Embodied Data Acquisition, Real Machine Deployment Now Has a Standardized Entry PointRoboSpeak — WeChat
- LeRobot Releases Cosmos3-Edge Policy Model for Robotic Post-TrainingLeRobot — Model repository updates
- SimReady Central: A Gateway to Simulation-Ready 3D Assets for Robot AI TrainingRobotics — Robot learning news discovery