Manipulation Latest coverage

LocoPilot: Single-Operator Bilateral Teleoperation for Loco-Manipulation Policy Learning

This paper presents LocoPilot, a single-operator bilateral teleoperation framework for loco-manipulation policy learning. It enables direct capture of robot-specific signals like proprioception and body-mounted observations without motion-capture instrumentation. The system integrates a 12-DoF quadruped with a 6-DoF manipulator, using two handheld controllers and a follower arm for force feedback. A ROS~2 pipeline records synchronized multimodal data and a unified 11D action space, validated through deployment on a physical robot without retargeting or additional inverse kinematics.

OpenAMRobot 2.0: open-source dual-arm mobile manipulator on ROS 2, rc.1 on 20 November

OpenAMRobot 2.0 is a general-purpose, open-source dual-arm mobile manipulator with an embodied AI software layer, built on ROS 2. It features two Enactic OpenArm 2.0 arms, a NVIDIA Jetson Orin NX compute module, and a safety layer with emergency stops. The robot is designed for tasks like grocery order picking and is being tested in Berlin, with a final release planned for December 2026.

KAIST Humanoid v0.9 Robot Pushes 77 Pounds and Lifts 44 While Walking

KAIST's Humanoid v0.9 uses deep reinforcement learning and onboard vision to coordinate walking with heavy loads, pushing 35 kg (77.2 lb) and lifting 20 kg (44.1 lb) while walking without external cameras. The demonstration highlights the challenge of whole-body locomanipulation, balancing forces during physical interaction with objects.

Ropedia Super Week Showcases Six New Research Projects Highlighting Approaches to Robot Learning and Control

Ropedia showcased six research projects at Ropedia Super Week, including two Foundry initiatives (EgoXperience and HoliWorld) for structuring egocentric human experience into data, and four studies on spatial reasoning, world modeling, and robot learning and control. The projects include datasets like EgoX-Data-6M with approximately six million frames, benchmarks showing performance improvements such as a 13.9-point score increase on MMSI-Bench for S-Agent 2.0, and simulation results indicating task success rates of 55% for model-guided action selection versus 26% for random selection. Ropedia aims to build data infrastructure connecting human experience with robot capabilities for Physical AI.

Eka Robotics' $1.45B Valuation Driven by Physical AI Model for Dexterous Manipulation

Eka Robotics, founded by MIT professor Pulkit Agrawal and former Google DeepMind researcher Tuomas Haarnoja, has raised $150 million in Series A funding, valuing the company at $1.45 billion. The startup's physical AI model, which uses a vision-force-action architecture, enables a robot to manipulate objects with high reliability and speed, surpassing human performance in some metrics. The model leverages simulation and reinforcement learning to scale training, though real-world deployment remains a challenge.

Robotics in Logistics: Sorting as the First Step

The article examines the deployment of logistics robots in warehouse sorting tasks, focusing on non-humanoid designs that are optimized for handling random package configurations. It discusses the technical and economic rationale for using these robots, including their adaptability to real-world conditions and the importance of real-world data for training embodied AI. The piece also explores potential future applications beyond sorting, such as in e-commerce fulfillment and warehouse operations.

Robotics in Logistics: The Rise of Non-Humanoid Sorting Robots

The article explores the deployment of non-humanoid robots in logistics, focusing on parcel sorting. It discusses the technical and economic advantages of these systems, including their ability to handle high-frequency, non-structured tasks, and the shift from demonstration to real-world deployment. Key players like StarDance, SelfVariable, and Jiajia Vision are highlighted, along with their approaches to automation in logistics scenarios.

DexOAK: Object-Aware Kinematic Retargeting

DexOAK is a learning-based framework that generates contact-consistent robot hand references from human motion-capture data. It uses an interaction-graph Laplacian objective to reproduce object-relative interaction geometry, and a feed-forward network conditioned on human keypoints and object point clouds to learn retargeting. Object-aware finetuning refines the network under contact, non-penetration, and self-collision losses, improving performance on tasks like OakInk-v2 and ARCTIC. The model transfers to unseen data and enables dexterous robot data generation from human data.

Climbing Robot Market Overview: Technical Evolution and Player Landscape in Extreme Throughput Scenarios

The article discusses the growth of the climbing robot market, driven by the need for high-throughput, space-efficient automation in warehouses. It compares key players such as Exotec, HaiPick Climb, ZeeBot, and RS Hyper, analyzing their technical specifications, system efficiency, and deployment strategies. The focus is on how these robots are evolving to meet the demands of high-density, high-speed warehouse operations.

Pulkit Agrawal@pulkitology· 19:02SignalEditorial score7878

This paper presents a breakthrough in robotic manipulation by showing that contact-based joint position signals can enable precise, zero-shot insertion tasks such as pegs, gears, and nuts without cameras or tactile sensors. The approach leverages joint drift to perform blind, fine-grained assembly tasks, marking a significant step toward general-purpose robotics.

The Robot I Didn’t Expect to Build

Jonathan W. Hurst, co-founder of Agility Robotics, discusses the development and launch of Digit 5, the company's first humanoid robot designed for industrial use. The robot is engineered to handle heavy warehouse tasks, reflecting a decade of research and development aimed at creating a commercially viable product. The article outlines the evolution from early prototypes like ATRIAS and Cassie to the current Digit 5, emphasizing the challenges of transitioning from research to real-world deployment.

A Reconfigurable Robotic System Using an Interbody Wire for Merged and Separated Tasks

Researchers from the University of Tokyo present TwinX, a system of two identical wheel-legged robots connected by an interbody wire. The robots can merge into different configurations (quadruped, swerve drive, or mobile manipulator) or operate separately, using the wire to perform tasks like garbage collection, object lifting, and slope traversal. The wire enables dynamic interaction and force transmission, offering new design possibilities beyond rigid joints.

2026 China International Industrial Fair Robotics Pavilion Preview: Robots Shift from Dancing to Competing for Workspaces

At the 2026 China International Industrial Fair, robotics exhibits shifted from performance-based demonstrations to practical industrial applications. Robots were shown performing tasks such as package sorting, chip insertion, and assembly line operations, highlighting a move toward real-world deployment and integration in manufacturing and logistics.

Two Unitree G1 Robots Play Table Tennis Doubles: PingpongDuo Enables Humanoid Team Coordination

A research team from Shanghai Chuangzhi College, Yuance Future, Hong Kong University, and NVIDIA has developed PingpongDuo, a system that enables two Unitree G1 humanoid robots to perform table tennis doubles. The system combines perception, team coordination, and whole-body motion control, using reinforcement learning for planning and risk assessment. Real-world testing showed improved performance over traditional rule-based strategies, with the robots achieving 18 consecutive alternating returns in a human-robot doubles match.

China's Robot Plays Badminton for 1,452 Consecutive Shots. Secures Guinness World Record, Boston Dynamics' Leader 'Hard Controlled'

The badminton robot developed by Zhejiang Shenchenkaidong Technology Co., Ltd. completed 1,452 consecutive shots in Shaoxing, earning a Guinness World Record certification. The robot, based on a self-developed general-purpose physical motion platform, achieves high-precision real-time perception, trajectory prediction, and coordinated control, showcasing engineering capabilities in dynamic competitive scenarios and demonstrating potential applications in logistics, warehouse automation, and other dynamic environments.

Wound-helix robot leaps and bounds over obstacles

Researchers from UCLA and the University of Michigan have developed a frog-like robot that uses flexible rods to snap between shapes, enabling it to hop over obstacles, climb steps, and swim. The robot outperforms rigid-legged designs in speed and adaptability across various surfaces, offering potential applications in search-and-rescue and inspection tasks.

After Collaborative Robot Deliveries Surpass 50,000 Units, This Company Brings Embodied Intelligence into Factories

At the 2026 China International Industrial Robot Expo, ABB Intelligent Robotics showcased its progress in deploying flexible manufacturing solutions, including a complete automotive production line with G-series and iS-series collaborative robots, as well as a new dual-arm robot and standardized welding workstations. The 50,000-unit delivery milestone marks a key step in its industrial deployment strategy.

ETH Zurich scientists turn WUJI robotic hand into self-propelled mobile manipulator

ETH Zurich researchers have developed a system where a WUJI robotic hand can crawl, recover from falls, and perform manipulation tasks without a separate mobility system. Using reinforcement learning, the hand learns to support its own weight and move using its fingers. The work was presented in a paper titled 'Fingers as Legs: Learning Self-Supported Locomotion and Manipulation with an Anthropomorphic Hand.'

15 Humanoid Robots Deployed in Manufacturing Line, Real-World Data Set Released

At the 2026 Sustainable Global Leaders Conference, Leju Robotics and Zhao Feng Machinery jointly launched an embodied intelligence solution for automotive parts loading. The 15 humanoid robots completed a full loading cycle in under 40 seconds with a 98% success rate, operating for over 1000 hours. The project also released a real-world data set for the OpenLET community to support further research and development in embodied intelligence.

Sim-to-Real Reinforcement Learning for Path-Safe Target Reaching in Robotic Harvesting of Hydroponic Lettuce

This study introduces PSA-PPO, a target-conditioned local reaching framework for robotic harvesting of hydroponic lettuce. Trained entirely in simulation, PSA-PPO achieves 94.7% reaching rate in physical testing, outperforming classical planners. The method integrates safety constraints and trajectory-based encoding to improve performance in constrained environments.