Skip to content
#Data and simulation
Oct 27Tue
  1. arXiv cs.RO58

    Sybil-Resilient Multi-Drone Exploration of Indoor and Confined Spaces

    The paper studies the problem of Sybil attacks during multi-drone collaborative exploration of buildings: a compromised drone forges identities, claims frontiers, and sends fabricated scans, causing real drones to miss areas and generating nonexistent walls in the map. The authors propose five defenses that rely only on the drones' own sensors and maps, including auditing map data, checking for a body's presence at an identity's claimed location, trusting only identities whose bodies have been observed, quarantining identities unobserved by any drone, and binding each observed body to a single identity.

  2. arXiv cs.RO64

    RoboLongR: Causally consistent rewards guide robots to progressively master long-horizon manipulation

    RoboLongR proposes an offline-to-online reward learning framework that uses causal consistency to align reward progress with the prerequisites supported by execution history, guiding policies from behavior cloning to online reinforcement learning. Pretrained on 88 RLBench tasks and Open-X data, the method achieves an average success rate of 31.8% on 18 long-horizon tasks excluded from reward pretraining, 9.2 percentage points above the strongest comparison baseline; on four real-robot tasks it achieves a 57.5% success rate, 32.5 percentage points above the comparison method.

  3. arXiv cs.RO72

    AdaDexTrack: Dynamic Modulation for Adaptive and Generalizable Dexterous Manipulation Tracking

    AdaDexTrack proposes a modulator-in-the-loop framework for language-conditioned dexterous manipulation tracking, using three types of feedback correction—reference modulation, object-latent modulation, and positional-target modulation—to mitigate synthetic reference noise, embodiment bias from human-to-robot retargeting, and error accumulation in fixed-reference tracking.

  4. arXiv cs.RO71

    RoboWorM: A Benchmark for Evaluating Motion Realism and Acceleration Effects in Embodied Video World Models

    The authors propose RoboWorM, a deployment-aware benchmark for robotic manipulation video world models that examines structural and spatiotemporal fidelity using hardware priors such as robot kinematic trees and fixed camera configurations. The benchmark collects 1,605 episodes across single-arm, bimanual, and humanoid morphologies, covering Markovian and long-horizon non-Markovian tasks under single-view and multi-view setups, and proposes two deterministic metrics that do not rely on vision-language models: KIVA and RIGIS.

  5. arXiv cs.RO75

    Scan2Instruction and PRPT-KS: Demonstration-free adaptation for language-guided navigation from a single scan

    The study proposes Scan2Instruction and PRPT-KS, which use a single iPhone scan to build a training environment and adapt a language-guided navigation policy without human demonstrations. Across five real environments, RealNav achieved an 85% real-robot success rate, compared with 49% for a baseline post-trained on human demonstrations; in unscanned areas of the same buildings, it still reached 82%, versus 41% for the baseline.

  6. arXiv cs.RO60

    RoboReason: Improving robotics-relevant VLM reasoning with scalable data generation

    RoboReason proposes constructing spatial and temporal ordering tasks from real robot stereo camera and wrist camera images to evaluate the robotics-relevant reasoning capabilities of vision-language models; experiments show that several advanced VLMs perform poorly on the ordering tasks. The authors accordingly propose a scalable autonomous training data generation pipeline, and supervised fine-tuning and reinforcement learning on this data both significantly improve model performance on spatiotemporal ordering and downstream robotics and visual reasoning tasks. The work was accepted to the EMNLP 2026 main conference.

  7. arXiv cs.RO63

    RoboJev-27B: An Open-Source Vision-Language Model for Judging Robot Task Success

    RoboJev-27B is an open-source vision-language model that takes a task instruction and 10 frames of a manipulation attempt as input and predicts whether the robot completed the task. It achieves 95.2% balanced accuracy and 0.993 AUROC on 268 held-out episodes of the same robot model (SO-101), and an AUROC of 0.890 on 459 episodes from unseen labs or robots. Median inference time is 0.26 seconds per episode on a single H100, and the model, inference code, and evaluation results for 20,628 predictions are released together.

  8. arXiv cs.RO64

    Open-World Law-to-Action: Using vision-language priors and touch to infer the motion laws of parts with hidden physics

    The authors propose the Open-World Law-to-Action (L2A) framework, in which a vision-language model proposes, for each movable part, its motion type, the terms of its law of motion, and possible locking factors; the part's law is then identified online via Bayesian linear regression with closed-form evidence, and touch can overturn the priors.

  9. arXiv cs.RO62

    EXE: Egocentric Cross-Embodiment Pretraining with Factorized Latent Actions for VLA Models

    EXE is an egocentric cross-embodiment pretraining framework that learns factorized latent action representations from human manipulation videos, separating shared manipulation priors from embodiment-specific signals. The authors state that this shared action space lets robot policies acquire generalizable manipulation semantics from diverse human demonstrations, and that the same pretraining can be used for both parallel grippers and dexterous hands. On RoboTwin 2.0, RoboCasa GR1, and real-world manipulation tasks, EXE achieves consistent improvements over baselines without pretraining.

  10. arXiv cs.RO62

    HA-DiPO: A Diffusion Policy Optimization Framework for Heterogeneous Robot Cooperative Transport

    The authors propose HA-DiPO, a hierarchical multi-agent reinforcement learning framework operating under centralized training with decentralized execution (CTDE) for heterogeneous legged robots cooperatively transporting an unfixed shared payload. At the high level, a centralized critic jointly trains each robot's independent coordination policy, and diffusion policies represent multimodal coordination responses; at the low level, embodiment-specific locomotion policies track the coordination commands. The policies are learned from scratch through reward-driven simulation interaction, and at execution they use only local observations with no inter-robot communication.

  11. arXiv cs.RO62

    WorldFailBench: Can World Models Understand Policy Failures?

    The authors propose WorldFailBench, which uses three tasks—failure reason, failure stage, and failure boundary—to test whether world models faithfully simulate policy failures. The benchmark covers manipulation tasks in the LIBERO and RoboTwin simulation environments, as well as real-world manipulation tasks with the G1 humanoid robot. Experiments on seven world models show clear gaps in their understanding of policy failures.

  12. arXiv cs.RO71

    MimicAll: A Structured Embodiment Conditioning Framework for Cross-Humanoid Whole-Body Motion Tracking

    MimicAll is a reinforcement learning framework that encodes global and joint-alignment conditions using structured embodiment descriptors, enabling heterogeneous humanoid robots to share a whole-body motion tracking policy. Training covers 4,096 variants across 16 humanoid robot families, and a benchmark of 464 bodies across five distribution strata was built. With the same amount of training transitions, shared training achieved a 99.4% 30-second success rate on randomized G1 bodies in 100STYLE, versus 42.9% for single-body specialized policies.

  13. arXiv cs.RO62

    ReSight: Task-Driven Self-Improving Memory for Embodied Agents

    ReSight is a task-driven self-improving memory system for embodied agents that retains keyframes and their spatiotemporal associations, and separates observation evidence from extensible derived fields so that historical observations can be reprocessed for new needs. Through Explore-Distill-Mount (EDM), the system identifies capability gaps from execution and exploration records and converts exploration-derived solutions into reusable memory builders, query tools, and skills.

  14. arXiv cs.RO67

    Survey: Dexterous hands, simulators, training, and evaluation in 113 learning-based manipulation methods

    A survey reviews 113 learning-based dexterous manipulation methods, covering the dexterous hands used, the training simulators, and how results are reported. The authors note that nearly half of the works use the same recipe: reinforcement learning on privileged state in a GPU simulator, then distillation into a vision-only student model transferred to hardware, and nearly all of them concentrate on the same simulator.

  16. arXiv cs.RO62

    B-Adapt and BVSLAM: Adaptive Multi-Exposure Bracketing for More Robust Visual SLAM

    The authors propose B-Adapt and BVSLAM to address image saturation and lost visual correspondences in visual SLAM when automatic exposure fails to adjust in time under rapid illumination changes. B-Adapt captures adaptive exposure brackets to cover the scene's dynamic range while maintaining photometric overlap, and BVSLAM uses the multi-exposure sequence directly as constraints in factor-graph state estimation without relying on HDR merging.

  17. AgiBot73

    AgiBot releases Genie Studio 2.0 embodied intelligence development and deployment platform

    AgiBot has released Genie Studio 2.0, an embodied intelligence application platform covering development, deployment, and operations. It consists of three modules—Flywheel, Agent, and Monitor—that connect data collection, model training, evaluation, inference, task orchestration, and robot fleet management.

    Why it matters: The original text provides the platform's three major modules and validation data from multiple production lines, which can be used to assess the engineering chain of embodied applications from development to operations.

  18. 江苏箸境智能科技有限公司33

    Zhujing's two projects selected for Jiangsu Physical AI high-quality dataset program; humanoid robots "Xiaozhu" and "Jingzai" appear at 2026 Physical AI and World Models Conference

    Two projects led by Zhujing were selected for Jiangsu Province's 2026 high-quality dataset program: the "High-Quality Dataset Construction Project for Embodied Intelligent Robot Logistics Sorting and Handling" was named a key construction project, and the "Embodied Intelligence Atomic Skill Dataset for Office, Catering and Other Scenarios" project was named a cultivation project. Zhujing brought its humanoid robots "Xiaozhu" and "Jingzai" to the exhibition floor of the 2026 Physical AI and World Models Conference. Its embodied intelligence datasets for four scenarios—office, supermarket, catering and housekeeping—have been listed and traded on the Jiangsu Province Data Exchange.

  19. 江苏箸境智能科技有限公司30

    Zhujing's two projects selected for Jiangsu Physical AI high-quality dataset program; humanoid robots "Xiaozhu" and "Jingzai" appear at 2026 Physical AI and World Models Conference

    Two projects led by Zhujing were selected for Jiangsu Province's 2026 high-quality dataset program: the "High-Quality Dataset Construction Project for Embodied Intelligent Robot Logistics Sorting and Handling" was named a key construction project, and the "Embodied Intelligence Atomic Skill Dataset for Office, Catering and Other Scenarios" project was named a cultivation project. Zhujing brought its humanoid robots "Xiaozhu" and "Jingzai" to the exhibition floor of the 2026 Physical AI and World Models Conference. Its embodied intelligence datasets for four scenarios—office, supermarket, catering and housekeeping—have been listed and traded on the Jiangsu Province Data Exchange.

  20. OFweek Robotics58

    Topstar's first three quarters net profit attributable to shareholders up 196% year on year, industrial robot revenue up 88.83%

    Topstar's 2026 third-quarter report shows that in the first three quarters, operating revenue was 1.969 billion yuan, up 16.68% year on year, and net profit attributable to shareholders was 146 million yuan, up 196.38% year on year. Revenue from industrial robots and automated application systems was 1.028 billion yuan, up 88.83% year on year; industrial robot orders totaled about 10,000 units and shipments about 8,600 units.

  21. OFweek Robotics42

    JPMorgan Report: Optimus Mass Production Approaches, Capital Sentiment and Real Data Gap Coexist

    A JPMorgan research report dated October 23 noted that Tesla's Optimus is approaching mass production at the Fremont factory, with a production guidance of 10,000 to 20,000 units. The recent rebound in Chinese complete robots is largely sentiment-driven, while actual demand remains R&D and testing procurement. At the Shanghai International Embodied Intelligence Skills Competition held the same day, 800 teams participated, and the competition's organizers stated outright that real-world scenario data collection is far from sufficient, limiting robots' generalization capabilities.

  22. 钛媒体60

    Xingyuan Intelligence launches cross-embodiment data engine Chorus Zhongxing

    On October 27, Xingyuan Intelligence launched Chorus Zhongxing, a cross-embodiment data engine that cleans and 3D-reconstructs human manipulation videos from multiple sources, including smart glasses, chest-worn cameras, and UMI devices, and adapts them to various robot embodiments. Under the current test configuration, Chorus achieves an effective data conversion rate of about 85% and adapts to 16 robot embodiments; one hour of pre-filtered human video can generate about 13.6 hours of cross-embodiment motion trajectories, and with 3x instruction augmentation it achieves about 40x effective training data expansion.

  23. Beijing Humanoid Robot Innovation Center58

    Beijing Humanoid launches Huisi Kaiwu App Store, robot skills downloadable and ready to use

    Beijing Humanoid launched the "Huisi Kaiwu App Store" in October. Robot skills are packaged as apps that can be searched, downloaded, and deployed with one click to run on full-size general-purpose robots and other bodies. The store also offers datasets and simulation assets, and the official skill app library is open for free; developers who list skills and have them called for a fee can share in the revenue. Users can search for "Jushen Tiangong" in mobile app markets and enter the app store as developers to use the relevant features.

  24. HuggingFace Daily Papers62

    QCP: Learning Quality-Controllable Robot Policies with Iterative Hindsight Relabeling

    The study proposes Quality-Controllable Policy (QCP), which enables inference-time control over individual quality attributes such as execution speed and motion smoothness, as well as their combinations. To extend the controllable quality range beyond the initial demonstrations, the authors designed an iterative hindsight relabeling loop: collect online rollouts under target quality labels, add successful trajectories to the dataset, then relabel the accumulated trajectories in hindsight according to their measured quality and fine-tune the policy.

  25. Robotics & Automation News58

    Comau integrates Nvidia technology to accelerate development of AI-driven robots

    Comau announced the integration of Nvidia technology into its product portfolio and design processes, including industrial robot controllers using Nvidia Jetson edge computing, connecting its robot product line to Nvidia Isaac Sim, and using the Nvidia software stack to optimize the vision capabilities and control strategies of its intralogistics products.

Oct 26Mon
  1. Guyuehome49

    NeuroMechFly: MuJoCo simulation of hexapod locomotion and sensorimotor closed loop in Drosophila

    A digital fruit fly reconstructed from micro-CT scans achieves hexapod walking in MuJoCo; the tripod gait self-organizes from phase-oscillator phase-locking, with reflex correction rules and adhesion measurements added. Its compound-eye vision outputs 721 ommatidia per eye, and all conclusions come from complete simulations with fixed seeds.

  2. 机器人全球资讯 · 微信公众号47

    Embodied AI robot mountain-climbing training ground built at foot of Mount Tai, replicating 3.8 km Mount Tai course to prepare for first CMG World Robot Mount Tai Climbing Competition

    Taiying Technology Group has built an embodied AI robot mountain-climbing training ground at the foot of Mount Tai, replicating the 3.8 km, more than 3,100-step terrain of the Mount Tai section from Hongmen to Zhongtianmen, with 11 test tracks for humanoid robots and quadruped robot dogs to test and tune for the first CMG World Robot Mount Tai Climbing Competition. The training ground includes special tracks such as low-recognition steps, steep, slippery, narrow steps, and S-shaped sharp-turn steps, focusing on assessing environmental perception, dynamic balance, and gait-switching capabilities. It has been selected as one of Shandong Province's first batch of key-supported regional hub training grounds.

  3. 未来机器人网 · 微信公众号58

    Nomura report assesses humanoid robots: brain not yet at GPT-2 stage, dexterous hands to scale first

    Nomura Securities released a research report on humanoid robots and China's embodied intelligence industry, arguing that robots are developing along the path of large language models, but their current brain has not yet reached the GPT-2 stage, and world model maturity is even lower than AI in 2012. The report proposes five test metrics: continuous working hours, fine hand manipulation, speed of adapting to new embodiments, number of demonstrations, and factory-floor accuracy. It says robots currently fall far short on all metrics, with factory-floor accuracy at about 60% versus the industrial requirement of over 98%.

  4. 机器人产业应用 · 微信公众号62

    NVIDIA Long-WAM and Meta RoboJEPA Submitted on the Same Day: Which Is Worth More, Long Memory or Scaling Laws?

    Long-WAM from teams at NVIDIA, MIT, the University of Hong Kong, and UCSD, and RoboJEPA from Meta FAIR, Mila, and other institutions were submitted to arXiv on the same day, October 7, measuring the returns from adding visual history and from expanding training compute, respectively.

  5. 36氪 · 微信公众号27

    Tsinghua AI4S Forum: Li Luming, Yaghi and others discuss how to use AI for scientific research

    The Tsinghua University Student AI4S Association held a forum on October 9. President Li Luming said bluntly that "a lot of research work is already meaningless," while Nobel laureate Omar Yaghi proposed "choosing science" and described an autonomous running laboratory in which AI proposes materials solutions and robotic systems synthesize and characterize them. Several professors stressed the building of foundational databases: Gao Xiangyu of Peking University Cancer Hospital advocated "one person, one file" with multi-point sampling throughout, Tsinghua's Xu Zhiping used physical transmission to connect simple and complex systems, and Sinopec's Sun Lili warned that data errors amplify step by step along the chain.

  6. 机器之心 · 微信公众号75

    COCO Matrix releases robot model CO01, using experience prompts to let robots learn while doing

    COCO Matrix released the robot model CO01, which treats demonstrations, failed attempts, and real-time guidance as experience prompts in context, allowing adaptation to new tasks without updating weights. Under the same pretrained model, a roughly 1-minute demonstration achieved an average success rate of 52% on unseen tasks; fine-tuning to a similar level required about 180 demonstrations and about 9 hours. In real-robot evaluations with at most five attempts per task, cumulative success rate rose from 33% on the first attempt to 84% when retaining its own history, versus 56% without retaining history.

  7. 机器之心 · 微信公众号78

    RoboCoach: Using World Models to Locate Weak Subtasks and Improve Long-Horizon Task Success Through Targeted Teaching

    Zhang Changshui and Cui Sen's team at Tsinghua University propose RoboCoach, an embodied intelligence system that uses the world model CoachWorld to run closed-loop trials of long-horizon tasks in imagination, identify the first weak subtask that times out, and request targeted human teaching for it, updating only the corresponding skill expert. Under a fixed demonstration budget, the success rate rises from 13.3% to 75.0% on Franka and from 40.0% to 83.8% on AgileX, while the uniform collection baseline reaches only 30.0% and 47.5%, respectively. On four unseen new task routes, the average success rate is 35.0%, compared with 0% for the shared-policy baseline.

    Why it matters: The paper presents a complete experiment on using a world model to locate weak subtasks and improve success rates through targeted teaching; its approach to data budget allocation may be a useful reference.

  8. 机器之心 · 微信公众号62

    UIUC and others propose DSReg, making each JEPA representation dimension correspond to a real-world latent variable

    Researchers from UIUC, Cold Spring Harbor Laboratory, Brown University, and the Max Planck Institute propose DSReg (Dependency-Sparsity Regularization), which can be added directly to an already-trained LeJEPA so that each learned dimension corresponds to a real-world variable, without reconstruction of observations, a decoder, or labels.

  10. 机器之心 · 微信公众号71

    Tsinghua and National University of Singapore Propose ParallelWorld: Multi-Step Rehearsal Improves Embodied Planning Success Rate

    Researchers from Tsinghua University and the National University of Singapore propose ParallelWorld, which adds multi-step candidate future rehearsal to high-level planning for embodied agents, raising task success rate from 56% to 83%. The method generates and verifies multiple candidate action branches in a rehearsal loop; during execution it takes only the first action of the selected trajectory, then replans based on new observations. Success rates were higher than the direct baseline across eight LIBERO-Pro settings, but adaptive search took on average about 5.8 times as long as direct execution.

  11. 硅星人Pro · 微信公众号67

    Hesai launches Kosmo, a 3D world data engine featuring the AI spatial camera Kosmo One and the Kosmo World cloud platform

    Hesai has launched Kosmo, a 3D world data engine consisting of the handheld AI spatial camera Kosmo One and the Kosmo World cloud platform. The former captures real environments, while the latter converts the captured results into editable, reusable 3D spatial data.

  12. 深蓝AI · 微信公众号75

    NVIDIA and Simon Fraser University propose GATOR: reconstructing posed 3D assets from casual photos

    A NVIDIA team, together with Simon Fraser University, proposes GATOR, which uses generative reconstruction as a base and multimodal agents for local refinement to reconstruct complete 3D objects with scene poses from casually taken photos. The method is trained only on synthetic scenes and, without any real-world fine-tuning, leads baselines on geometry, appearance, and pose metrics in real tabletop and indoor scans; without refinement, it takes about 0.5 minutes to outperform GPT-6-Astra running for 20 minutes. The authors also note that reliance on synthetic data, binding to commercial APIs, and stability under extreme poses remain open issues for deployment.

  13. Xbotics具身智能实验室 · 微信公众号62

    Xbotics Lecture 6, Part 2: The full pipeline from object pose to grasp pose, with hand-eye calibration practice

    The Xbotics Embodied Intelligence Lab has released Part 2 of Lecture 6 in its "Twenty-One Lectures" series, explaining how to generate grasp candidates from object poses and complete the coordinate transformation from the camera frame to the robot base frame. The article breaks down five elements: grasp point, approach direction, opening width, safety offset, and orientation matching. It compares the Eye-in-Hand and Eye-to-Hand hand-eye calibration topologies and provides procedures for reachability checks, degradation under observation failure, and troubleshooting perception failures.

  14. Robohub27

    September 2026 robotics society, business, and culture roundup: Global industrial robot stock tops 5 million units, Toyota plans to invest 1 trillion yen, ETH Zurich achieves humanoid hand walking

    The International Federation of Robotics' "World Robotics 2026" report shows that more than 5 million industrial robots are now in operation in factories worldwide; in 2025, more than 600,000 units were newly installed, up 11% year on year, with China accounting for 354,000 units, or 59% of the global total.