Robotics Research Publications

Track recent robotics research publications, benchmarks, datasets, model releases, and technical papers. Browse dated, source-backed robotics company events.

Verified Events

Sorted by event date, newest first.

Browse the company directory for individual company profiles and the events recorded against each one.

Research PublicationAstribot

Astribot releases SmoothRL for online learning on S1

Humanoid

· research publication · Astribot · Astribot S1 · Humanoid

Evidence notes
  • SmoothRL separates committed, executed and discarded actions so policy gradients use only actions that physically ran.
  • After 250 episodes on S1, reported success reached 94% for dynamic tossing, 83% for pen capping and 90% for box opening.
  • The training loop supports VR takeover and residual joystick correction while the robot continues operating.
Date
Company Astribot
Product Astribot S1
Event type Research Publication
Segment Humanoid
Added to Korthos
Research PublicationAxis Robotics

Axis Robotics releases Axis Sim Dataset V1

Data Infrastructure

· research publication · Axis Robotics · Data Infrastructure

Evidence notes
  • The public release includes the dataset, training code and benchmarks for a simulated Franka Research 3 arm.
  • Continual pretraining on V1 increased π0.5 success on LIBERO-Plus from 83.9% to 88.8%.
  • The trajectories cover pick-and-place, stacking, pouring, articulated-object manipulation and tool use collected through Axis Hub.
Date
Event type Research Publication
Segment Data Infrastructure
Added to Korthos
Research PublicationENGINEAI

ETH Zurich study demonstrates autonomous sparse-structure traversal on PM01

Humanoid

· research publication · ENGINEAI · PM01 · Humanoid

Evidence notes
  • The PM01 completed the full monkey-bar sequence in 14 of 15 hardware trials across three bar configurations and reached brachiation speeds up to 0.5 m/s.
  • A separately trained policy completed ten passes beneath randomly placed 2 cm wooden bars without contact, recovering to a stable stance after each crossing.
  • The research configuration replaced PM01 hands with passive hooks and mounted a RoboSense E1R lidar; after one goal position was supplied, the traversal sequence ran autonomously.
Date
Company ENGINEAI
Product PM01
Event type Research Publication
Segment Humanoid
Added to Korthos
Research PublicationAlibaba Group

Amap releases open ABot-Recon streaming 3D model

Foundation Models

· research publication · Alibaba Group · Foundation Models

Evidence notes
  • ABot-Recon accepts monocular RGB video without requiring depth sensors or pre-calibrated camera parameters.
  • On KITTI-02 at 504 by 280 resolution, ABot-Recon reaches 24.45 frames per second with about 6.71 GB peak memory; the stated minimum GPU is a GTX 1080 Ti.
  • Inference code, evaluation scripts and pretrained weights were public at announcement; no future training-code release is treated as completed.
Date
Event type Research Publication
Segment Foundation Models
Added to Korthos
Research PublicationX Square Robot

X Square introduces WALL-SS long-horizon robotic world model

Humanoid

· research publication · X Square Robot · Humanoid

Evidence notes
  • WALL-SS generated action-conditioned streaming rollouts up to 60 seconds while retaining recent history at fine resolution and compressing older state under a bounded memory budget.
  • Across 600 matched simulated and physical rollouts on six manipulation tasks, task-level success rates had mean absolute error of 0.062 and Pearson correlation of 0.93.
  • A co-trained action expert controlled a physical bimanual platform across seven tabletop tasks and reached average Task Progress of 69.1, ahead of the three reported comparison policies.
Date
Event type Research Publication
Segment Humanoid
Added to Korthos
Research PublicationNexdata

Nexdata releases multimodal egocentric dataset suite for Physical AI

Data Infrastructure

· research publication · Nexdata · Data Infrastructure

Evidence notes
  • Modalities include video, IMU, SLAM and depth data, hand keypoints and semantic annotations.
  • The suite is designed to support foundation-model training through fine manipulation and embodied task learning.
  • Nexdata offers global collection alongside hardware procurement, annotation, preprocessing and quality validation.
Date
Company Nexdata
Event type Research Publication
Segment Data Infrastructure
Added to Korthos
Research PublicationPerceptron

Perceptron releases open Isaac 0.5 embodied foundation model

Foundation Models

· research publication · Perceptron · Foundation Models

Evidence notes
  • Inputs can include images, video, language, robot state and previous actions; outputs include grounded visual responses, task progress and robot actions.
  • Perceptron reports training across more than 35 robot systems, 100,000 hours of robot experience, one million hours of general video and three trillion multimodal tokens.
  • Weights, training and inference code, the technical report and LeRobot support were released publicly.
Date
Company Perceptron
Event type Research Publication
Segment Foundation Models
Added to Korthos
Research PublicationLightwheel

Lightwheel releases EgoSuite-Open100K egocentric human dataset

Data Infrastructure

· research publication · Lightwheel · Data Infrastructure

Evidence notes
  • The collection spans seven environment categories, 128 scene classes, more than 15,000 collection scenes and more than 15,000 tasks.
  • Head- and wrist-mounted views are paired with depth, hand-pose, full-body-pose and event-level semantic annotations for manipulation, coordination and task-sequence learning.
  • The first tranche is distributed through Hugging Face and AtomGit, and Lightwheel said the project will be donated to the OpenAtom Foundation for incubation.
Date
Company Lightwheel
Event type Research Publication
Segment Data Infrastructure
Added to Korthos
Research PublicationGeneralist

Generalist releases GEN-1.5 with one-shot physical-task learning

Foundation Models

· research publication · Generalist · GEN-1 · Foundation Models

Evidence notes
  • Across 10 tasks, Generalist reports 59% average success from one 3–12 second in-context demonstration and 83% after 10 gradient steps on five minutes of task data.
  • The multimodal model maintains 30 seconds of video context alongside sensor, language and proprioceptive inputs, and produces action trajectories at 100 Hz.
  • The release also shows prompts transferring from simulation to a physical robot and, in some cases, from human-hand demonstrations to immediate robot execution.
Date
Company Generalist
Product GEN-1
Event type Research Publication
Segment Foundation Models
Added to Korthos
Research PublicationNoitom

Noitom releases the HiPHI motion and interaction dataset

Data Infrastructure

· research publication · Noitom · Data Infrastructure

Evidence notes
  • The dataset contains recordings from 132 performers, including 371.8 hours of motion and 245.7 hours of interaction data.
  • Noitom reports 90 Hz capture across 40 objects in 12 categories.
  • Policies trained with HiPHI were evaluated through transfer to a physical Unitree G1; commercial licensing is offered separately.
Date
Company Noitom
Event type Research Publication
Segment Data Infrastructure
Added to Korthos
Research PublicationSkild AI

Skild AI publishes S1 in-context robot learning model

Foundation Models

· research publication · Skild AI · S1 · Foundation Models

Evidence notes
  • S1 applies in-context learning to seen and unseen manipulation tasks, including plant potting, pancake preparation, pour-over coffee and kit assembly.
  • Skild reports unseen tasks lasting up to 10 minutes and a sevenfold improvement over language-only prompting on its evaluation set.
  • The official release demonstrates transfer from a human video example to robot execution, including an 11-minute demonstration-to-execution plant-potting workflow.
Date
Company Skild AI
Product S1
Event type Research Publication
Segment Foundation Models
Added to Korthos
Research PublicationDyna Robotics

Dyna Robotics details the million-hour infrastructure behind Dyna-2

Industrial

· research publication · Dyna Robotics · Industrial

Evidence notes
  • A new episode format with topic-group chunking reduced storage by about 68% and made sample reads approximately 2.9 times faster.
  • An Airflow-based ingestion architecture increased throughput from 14,000 to 440,000 episode-hours per week, cutting million-hour processing from roughly 16 months to under three weeks.
  • Warehouse-backed curation and memory-mapped tables reduced time to first training batch from about 48 hours to under one minute across a 43-million-episode dataset.
Date
Event type Research Publication
Segment Industrial
Added to Korthos