Unitree G1
PulseAugur coverage of Unitree G1 — every cluster mentioning Unitree G1 across labs, papers, and developer communities, ranked by signal.
- 2026-07-09 research_milestone Humanoid robots controlled by surgeons performed a world-first surgery on live pigs. source
- 2026-07-09 research_milestone A Unitree G1 humanoid robot, controlled by surgeons, performed a gallbladder removal surgery on live pigs, marking a significant advancement in remote robotic surgery. source
- 2026-07-05 controversy A Unitree G1 robot reportedly injured office workers in Indonesia. source
- 2026-06-20 research_milestone A Unitree G1 humanoid robot successfully summited Ecuador's Chimborazo volcano at 6,310 meters. source
17 day(s) with sentiment data
-
Humanoid robot navigates confined spaces using RL-based walking and crawling
Researchers have developed REACH, a novel reinforcement learning framework enabling the Unitree G1 humanoid robot to navigate confined spaces by seamlessly switching between bipedal walking and crawling. This hierarchic…
-
ActiveAlign framework uses active perception to bridge real-world and simulation gaps
Researchers have introduced ActiveAlign, a novel framework designed to bridge the gap between real-world robotic interactions and their simulated counterparts. This system utilizes active perception, mimicking human beh…
-
New framework enhances explainability for robot navigation decisions
Researchers have developed a new stage-aware attribution framework to improve the explainability of Vision-Language Navigation (VLN) models. This framework addresses the challenge of evidence distribution across histori…
-
New framework HumanoidNDM bridges robot training-deployment gap
Researchers have developed HumanoidNDM, a novel framework designed to bridge the gap between training and real-world deployment for humanoid robot motion tracking. This system utilizes a neural dynamics model trained in…
-
ViT-H framework enables humanoid robot control from video with motion tracking feedback
Researchers have developed ViT-H, a novel framework for controlling humanoid robots using video input. This system utilizes a diffusion model to directly generate humanoid reference motions from videos, bypassing interm…
-
HuGe models enable humanoid robots to perform whole-body control tasks
Researchers have developed HuGe, a new family of vision-language-action models designed for whole-body control in humanoid robots. These models are pre-trained on over 2,000 hours of egocentric data from both humans and…
-
Amap open-sources Lambda-0 humanoid model trained on first-person data
AutoNavi Holdings Limited, also known as Amap, has open-sourced Lambda-0, a humanoid foundation model. This model was trained using 500 hours of first-person human data. Lambda-0 has demonstrated the ability to perform …
-
New method speeds up humanoid robot training by predicting footsteps
Researchers have developed a new method for humanoid robot locomotion that significantly improves training efficiency and generalization. By using a lightweight prediction model to anticipate future footsteps, the syste…
-
Menlo Research releases open-source humanoid robot Asimov 1 for AI agents
Menlo Research has developed an open-source humanoid robot called Asimov 1, designed to serve as a physical embodiment for AI agents. The project aims to create a more accessible and repairable robot platform compared t…
-
New framework enables humanoid robots to interact physically with humans
Researchers have developed REACT, a framework designed to enable humanoid robots to engage in complex physical interactions with humans. This system utilizes vision-language models (VLMs) to interpret and transfer seman…
-
AI memory systems advance with new tools and research on consolidation
Recent developments in AI memory systems highlight advancements in both practical implementations and theoretical frameworks. The llama.cpp project has released version b11806, introducing support for prism.hadamard fol…
-
Humanoid robots gain parkour skills with new perception and training frameworks
Researchers have developed two new frameworks for enabling humanoid robots to navigate complex, unstructured terrains at high speeds. The first, "Hiking in the Wild," uses a perceptive parkour framework that maps depth …
-
HumanVerse dataset enables humanoid robots to learn loco-manipulation from human data
Researchers have introduced HumanVerse, a 500-hour dataset that pairs egocentric video with synchronized body and hand motion. This dataset is designed to train a humanoid foundation model for loco-manipulation tasks by…
-
Humanoid robots learn agile skills from human videos with new HumanX framework
Researchers have developed HumanX, a new framework designed to enable humanoid robots to learn agile and generalizable interaction skills from human videos. The system integrates XGen, a data generation pipeline that cr…
-
Hive Robotics trains Unitree G1 for autonomous supermarket restocking
Hive Robotics is training the Unitree G1 robot in Denmark to autonomously restock supermarket shelves. The company aims for the robot to achieve this capability by early 2027. This development highlights advancements in…
-
New Contextual Safety Filtering enhances robot motion generator safety
Researchers have developed a new method called Contextual Safety Filtering (CSF) to improve the safety of text-conditioned motion generators. CSF is a training-free filter that uses safe and unsafe reference trajectorie…
-
New FAITH framework enhances safe reinforcement learning for robots
Researchers have developed FAITH, a novel framework for safe reinforcement learning in high-dimensional systems. This model-free approach separates safety and task performance objectives, using a learned safety value to…
-
New framework ensures traceable and reproducible world states for robots
Researchers have developed Traceable World State (TWS), a new framework for robotic systems that focuses on maintaining a verifiable and reproducible world state. Unlike conventional methods that prioritize the latest e…
-
Robots learn dexterous manipulation from human play and videos · 10 sources tracked
Multiple research papers explore advancements in dexterous robotic manipulation, focusing on learning from human demonstrations and improving control strategies. Techniques like "Dexterity from Supervised Play" (DSP) an…
-
Tsinghua University's LATENT Robot Wins Award for Humanoid Tennis Forehand
A humanoid robot named LATENT, developed by Galbot and Tsinghua University, has received an IROS 2026 award for its impressive tennis forehand capabilities. The robot demonstrated a high success rate, returning forehand…