Skip to content
#Human-robot interaction and collaboration
Oct 22Thu
  1. arXiv cs.RO45

    An Explainability Framework for LLM-Based Assistive Robot Obstacle Avoidance Navigation

    The study proposes the Context-Aware Navigation Explanation framework, which uses a large language model combined with perceptual context and planned trajectories to generate natural-language explanations for assistive robot obstacle avoidance and navigation. The system is deployed on a Unitree Go2 platform in indoor navigation scenarios to evaluate how different explanation context configurations affect user trust and explanation satisfaction. Results show a strong positive correlation between explanation satisfaction and trust, and that the configuration combining obstacle information with future trajectory produced the best explanations.

  3. arXiv cs.RO46

    Sensorless force/torque observer and hybrid pose-force control framework for physical human-robot interaction

    The framework combines a neural-network-enhanced robust wrench observer with a hybrid pose-force controller for physical human-robot interaction (pHRI), enabling end-effector pose and force tracking without force/torque sensors. The neural network compensates for system friction and unmodeled dynamics to improve observer performance, and task-null-space decomposition separates joint velocities and reduces the effect of human disturbance on the observer. Experimental results show that the framework achieves robust and accurate wrench estimation while maintaining reliable pose and force tracking, and can be applied to various robotic arm platforms without force/torque sensors.

  4. arXiv cs.RO44

    Meta-Learning-Based Mode Recognition and Adaptive Control for Human-Robot Interaction

    The framework uses a meta-learning-trained neural network to recognize two human-robot interaction modes, active guidance and passive following, without manually setting switching thresholds. The passive mode recognition rate rose from 90.1% to 99.1%, and the active mode from 79.8% to 97.1%. An adaptive meta-learning controller can compensate for human disturbances and adapt online to varying interaction forces. A multi-trajectory meta-learning algorithm jointly learns an environment basis representation and force-related parameters to handle unmodeled dynamics. Experiments verified its robust control and reliable mode recognition performance.

  5. Droidup/卓益得43

    Moya appears as "Cuicui" at Chadong scenic area in Border Town, Xiangxi, serving as a tour guide

    Moya, the world's first fully biomimetic robot developed by Zhuoyide, took on the literary character "Cuicui" and completed a virtual-physical linked performance combining a "digital human + physical robot" on October 18 at the 5th Xiangxi Tujia and Miao Autonomous Prefecture Tourism Development Conference. The next day, it "reported for duty" as a tour guide at the Qule Visitor Service Center in the Chadong scenic area of Border Town, interacting with tourists.

  6. Droidup/卓益得42

    DroidUp's Moya Opens Guiyang Square Dance Competition and Serves as Anti-Fraud Promoter

    Moya, a fully bionic robot developed by DroidUp, delivered the opening announcement at the finals of the Huaguoyuan Square Dance Competition in Nanming District, Guiyang, and then explained anti-fraud knowledge to community residents in the government publicity interaction area with real-time voice interaction. Moya weighs about 30 kilograms, uses tendon-rope-driven compliant motion, has a touch close to body temperature, and is equipped with the self-developed DroidSoul full-body interactive agent, which can recognize expressions, tone, and posture and remember each interaction. DroidUp plans to bring Moya to near-human service scenarios such as high-end reception, health and elderly care companionship, educational growth, and entertainment IP.

  7. OFweek Robotics58

    AgiBot launches seven consecutive days of technology releases and will conduct a 24-hour convenience store livestream validation

    AgiBot announced that starting October 22, it will release seven full-stack technology achievements over seven consecutive working days, covering datasets, an operation platform, the heavy-duty robot G2 Max, the embodied Agent framework GenieMaster, a simulation platform, a foundation model, and a collective experience learning paradigm.

  10. Hyundai Motor Group28

    Hyundai Motor Group and RISD enter seventh year of collaboration with "Motion Rewild" research on embodied AI mobility and biomimetic design

    The collaboration between Hyundai Motor Group and the Rhode Island School of Design (RISD) enters its seventh year. The 2026–27 Regeneration Studio, themed "Motion Rewild," studies how nature moves in uncontrolled environments (such as disaster sites) and applies this to research on physical AI mobility and biomimetic design.

  11. HuggingFace Daily Papers58

    HiSTEVE 1.0: A Hierarchical VLA Agent for Open-World Games

    HiSTEVE 1.0 is a hierarchical VLA agent that achieves long-horizon autonomy and native keyboard-and-mouse control in Minecraft through a capability-aware agent harness. Its orchestration layer uses structured memory to track progress, with a VLM planner selecting subtasks within the VLA's capabilities; the control layer uses a single VLA to execute all subtasks with vision-only perception.

Oct 21Wed
  2. 36Kr Embodied Intelligence25

    In the AI Era, the Value of "People" Matters Even More in Later Life

    On the eve of the Double Ninth Festival, Beike Charity and People magazine held a salon on later life and released a companion booklet, "New World, Old Friends: Observations on Later Life in the AI Era." The salon invited eight guests, including Dai Jun, Hu Yong, Zhang Yue, and Cao Cuifen, to discuss aging amid the AI wave. Cao Cuifen said she would prefer a privacy-protecting robot as a caregiver, while Hu Yong argued that technology is only a supporting tool above the waterline of the caregiving iceberg.

  4. arXiv cs.RO62

    REACT: VLM-Guided Contact Retargeting and Generative Partner Modeling for Humanoid-Human Interaction

    REACT is a framework for learning deployable humanoid-human interaction policies from human-human demonstrations, using semantic contact structure as the supervisory signal that links demonstration processing, interaction learning, and onboard control. It uses a VLM-guided contact modeling and retargeting pipeline to identify semantic contacts and transfer them onto humanoid-human geometry, then distills a privileged interaction policy into an egocentric depth policy that relies entirely on onboard sensing, and introduces a generative partner state modeling objective to predict partner motion from occluded viewpoints.

  5. arXiv cs.RO57

    PointHand user study: How keypoint hand rendering affects novice teleoperation demonstrations and diffusion policy success rates

    The paper proposes the PointHand teleoperation interface, which renders a multi-fingered robotic hand as sparse keypoints at the fingertips and joints, matching the keypoints observable by the policy. In a simulated mug-in-box task, 18 novices and 4 experts collected 2,200 trials across three rendering conditions; among them, novices who saw emphasized rendering of the thumb and index finger achieved 441 successful demonstrations out of 600 trials, higher than the 320 out of the real hand-shape rendering.

  6. arXiv cs.RO48

    Evaluating Social Reasoning of Gemini Robotics-ER on the SHREC Benchmark

    An evaluation tested the social reasoning capabilities of Gemini Robotics-ER on the SHREC human-robot social interaction benchmark and compared it with published general-purpose vision-language model baselines. Results show that Gemini Robotics-ER outperforms GPT-4o-mini on error/competence detection but performs worse on fine-grained social-attribute judgments; adding a visual frame helped one task while hurting the other.

  7. arXiv cs.RO38

    Determination of the Human-Robot Interaction Workspace for Head Stabilization Based on a 6-DoF Stewart Platform Robot

    A study conducted human volunteer experiments using a 6-degree-of-freedom (6DoF) Stewart platform robot to determine the biokinematic workspace of the head-neck system, establishing foundational data for human-robot interaction in precision head stabilization. The system integrates real-time optical motion tracking and can dynamically compensate for head movement without physical restraint, for use in precision-demanding medical scenarios such as neuroimaging and intracranial radiotherapy.

  8. arXiv cs.RO58

    XoNav: Using a Hip Exoskeleton as a Personalized Haptic Interface for Assisted Navigation

    XoNav is a four-actuator hip exoskeleton that renders navigation commands directly as haptic cues at the hip joints, using sagittal-plane actuators to render an Advance cue and two-level Braking cues, and frontal-plane actuators to render two-level Steering cues. A bilateral thigh-mounted IMU setup combined with an LSTM gait-state classifier synchronizes cues with the wearer's gait and suppresses them while standing; the classifier achieved 95.2-97.9% recall on unseen users, with mean detection delays of 38.6 ms and 63.1 ms for stance-to-swing and swing-to-stance transitions, respectively.

  9. LeadeRobot65

    AgiBot Expedition A3 Ultra Full-Size Humanoid Robot Smart Retail Warehouse Opens at Wuhan Riverbank

    AgiBot has partnered with Wuhan Tourism Group and Zhongyun Weixun to deploy the Expedition A3 Ultra full-size bipedal humanoid robot for real-world commercial operations at the "Jiangyu" bridge-pier film camp on the Hankou riverbank. As of October 19, the Expedition A3 Ultra had run stably for 18 days, completed 827 orders cumulatively, and achieved an overall grasp success rate of over 97%; this weekend, single-day orders exceeded 100 with a grasp success rate of 100%.

  10. LeadeRobot36

    ArcVideo Technology proposes an "autonomous operation plus remote takeover" path for robots entering workshops

    At the 100th session of the Origin Space seminar, ArcVideo Technology Chairman Sun Yanlong proposed that for robots to enter real industrial scenarios, they must first overcome smoke, dust, occlusion, strong light, and unstable networks. The company is exploring an "autonomous operation plus remote takeover" path, using industrial inspection as an example, where robots autonomously complete routine tasks and remote operators intervene to handle anomalies.

  12. Ken Goldberg25

    RoboBusiness 2026 robotics summit lineup announced

    Rodney Brooks, Jonathan Hurst, and Ken Goldberg will share the stage to open RoboBusiness 2026, kicking off the "Next 20 Years" keynote. 1,600 attendees, 40+ sessions across 5 tracks, covering humanoid and field robotics, with speakers from Agility, Boston Dynamics, Tesla, Amazon Robotics, ABB, and others. The closing keynote on day one, "Beyond Demos: Robot AI in Production," is worth watching—moving from controlled demos to reproducible production deployment is the real challenge today. Ken Goldberg commented: Looking forward to this conversation!

Oct 20Tue
  1. 月泉仿生 · 微信公众号24

    ICBE 2026 Delegation Visits ISBE Beijing Exhibition Hall to Deepen International Exchange in Bionic Engineering

    On October 19, a delegation attending ICBE 2026, led by Professor Thomas Speck of the University of Freiburg, Germany, visited the International Society of Bionic Engineering (ISBE) Beijing Exhibition Hall and observed live debugging and dynamic demonstrations of the Y-Hand M2 bionic dexterous hand and a humanoid robot in the testing room. The exhibition hall, led by the ISBE Science Popularization Committee and funded by Yuequan Bionic, covers about 500 square meters and features five sections: Materials and Structures, Perception and Cognition, Actuation and Movement, Information and Intelligence, and Bionic Applications. It officially opened on July 26, 2025.

  2. Italian Institute of Technology22

    IIT and Fondazione Luzzati-Teatro della Tosse present three events at the Science Festival exploring the "Corpi aumentati" project and bionic prostheses

    The Italian Institute of Technology (IIT) and Fondazione Luzzati-Teatro della Tosse have partnered to organize three events for the Science Festival, centered on the IIT-coordinated "Corpi augmentati" project, exploring user-centered robotic prosthesis design.

  3. Italian Institute of Technology12

    Italian Institute of Technology (IIT) to take part in 2026 Genoa Science Festival

    The Italian Institute of Technology (IIT) will take part in the 24th Science Festival in Genoa, held from October 22 to November 1 under the theme "Prospettive," with 254 events in total. IIT's participation includes an exhibition on optogenetics, workshops on bionic prosthetics and body enhancement, and an online robotics special event, "Prospettive robotiche," led by Alessandra Sciutti.

  4. 雷科技55

    DJI, Insta360, Qiyuan, and Obsbot Double Down on AI Imaging: Will Photography Robots Be the Entry Point to Embodied Intelligence?

    The article reviews the AI imaging and photography robot initiatives of manufacturers including DJI, Insta360, Qiyuan, and Obsbot, and notes that photography robots are emerging at the intersection of two paths: the autonomization of imaging devices and the deployment of consumer-grade robots.

  5. OFweek Robotics44

    Musk criticizes humanoid robot industry: pre-choreographed demos are not real AI capability

    At the X Takeover event at Tesla's Texas factory, Musk criticized the humanoid robot industry, saying many complex demonstrations rely on pre-choreography and "that is not real AI"; the criterion should be whether a robot can handle situations it has never seen before. In the first half of 2026, global humanoid robot shipments exceeded 22,000 units, up nearly 300% year on year, but entertainment performances and education/science data collection together accounted for over 60%, smart manufacturing for only 13%, and warehouse logistics for only 5%.

  8. IEEE RAS54

    Agility Robotics co-founder Damion Shelton on the path to humanoid robot commercialization

    IEEE RAS interviewed Damion Shelton, co-founder of Agility Robotics and executive director of CMU's Swartz Center, who recounted how Digit chose its first commercial scenario: it once worked with Ford to explore last-mile delivery, but abandoned that because humanoid robots depend on autonomous driving, another deep-tech challenge, and instead focused on in-warehouse material handling in closed environments.

  10. 禾川科技 · 微信公众号42

    Hechuan Technology wraps up 2026 Shanghai Industrial Fair: full-stack industrial control solutions and FR1 industrial assembly robot on display

    At the 2026 Shanghai Industrial Fair, Hechuan Technology showcased its full-stack industrial control product matrix and debuted Sysctrl Studio, an AI programming assistant. The tool is natively integrated into domestic PLC programming software and supports natural-language PLC program generation, syntax checking, code optimization, and conversational troubleshooting, and is compatible with Hechuan's full controller lineup.

Oct 19Mon
  1. 小鹏机器人57

    He Xiaopeng announces XPENG's open cooperation across smart vehicles and physical AI at Paris Automotive Summit

    As the only Chinese automaker CEO invited to the Paris Automotive Summit, He Xiaopeng formally introduced XPENG's smart vehicle solutions business "XPENG AI ENGINEERING," opening three major areas to global partners: vehicle platforms, modular technologies, and toolchains. XPENG is also opening physical AI ecosystem technology cooperation covering humanoid robots, Robotaxi, flying cars, Turing AI chips, and physical-world foundation models.

  4. 月泉仿生 · 微信公众号34

    Yuequan Bionic Showcases Y-Hand M2 Dexterous Hand and W-Bot 2.0 at 2026 China Outdoor Sports Industry Conference

    At the 2026 China Outdoor Sports Industry Conference, Yuequan Bionic is exhibiting its bionic dexterous hand, the Y-Hand M2, and its wheeled humanoid robot, the W-Bot 2.0, at Booth B3 of the Beijing Changping venue and at the "Shanshui Changping" viewing point, demonstrating outdoor human-robot interaction.

  5. arXiv cs.RO55

    Benchmarking Service Robots in Public Environments: From the Lab to Real-World Operation

    This study proposes an interdisciplinary process for benchmarking service robots in public everyday environments, addressing the gap between laboratory metrics and real-world operation, with evaluation conducted at technology readiness levels 6-7. Over the three years 2023-2025, seven benchmarking events (one consensus workshop and six on-site evaluations) assessed three robots across three use cases: outdoor park cleaning, pedestrian underpass cleaning, and interactive assistance in a public library.