Addressing the issue of loud thuds when a humanoid robot lands, the investigation points to the difference in choosing between foot force control and position control. The article presents the root-cause diagnosis process for this abnormal noise in the form of a troubleshooting record, for developers to compare against their own control schemes.
Robots from about 90 teams competed in the first World Mountain Robot Challenge on Mount Tai, on a 3.9-kilometer course with over 600 meters of elevation gain, requiring them to race or carry fragile items such as eggs and watermelons to the summit. A quadruped robot from Beijing Lingyi Technology (a Unitree subsidiary) won the quadruped speed category in 43 minutes 41 seconds, while a humanoid robot from the Beijing Zhishen Lingyue team completed a shorter course in 27 minutes 26 seconds. The event was hosted by China Media Group and included awards such as Fastest Robot and Best Load-Bearing Quadruped Robot.
The paper proposes a Transformer-based architecture with shared modular recurrence for universal morphology robot control and evaluates its generalization performance on a set of MuJoCo robots. The authors note that contextual features are often only partially available but can be recovered through modular interactions, improving multi-robot control and generalization to contexts unseen during training. Results show substantial improvement in zero-shot generalization performance on robots with unseen dynamics, kinematics, and topologies across four environments.
The paper proposes actuator-aware (AA) observer inputs: the altitude and attitude observers use limited virtual inputs filtered by an assumed rotor model and aligned with the measurement delay, while the horizontal observers use acceleration reconstructed from measured attitude.
The research team proposes I3L, a hardware and software system that combines whole-body correction with iterative policy refinement to mitigate the out-of-distribution degradation of offline imitation learning policies in long-horizon whole-body mobile manipulation.
This paper releases an open-source ROS 2 software stack that restores ROS 2 support for the Franka Emika Panda and resolves the long-standing unreliability of its external position control interface. The authors' analysis attributes vibrations and protective stops to external control loop timing and sampling jitter rather than to limitations of the robot itself, and on that basis they propose an asynchronous hardware interface that decouples real-time communication from the ROS 2 control loop, a rate-matching mechanism for slower command sources, and a position-domain reference generation strategy.
This research uses manipulator motion for propellant-free attitude control of free-floating spacecraft, modeling it as a trajectory optimization problem with joint and collision-avoidance constraints and solving it with an interior-point method. Across a range of spacecraft-manipulator systems, the method computed rest-to-rest slew trajectories of up to 90° and detumble trajectories with a 40% reduction in angular velocity. The authors also directly compare the manipulator's control authority with that of reaction wheel arrays via momentum and torque envelopes, noting that manipulators could serve as a redundant or even primary attitude control system, with particular promise for in-space assembly and manufacturing involving the capture of high relative-mass payloads.
This study proposes Gaussian Evolutionary Specialists (GES), a reinforcement learning-based co-design framework. By maintaining a population of specialist policies whose design-space territories evolve according to their control performance, it mitigates behavioral diversity collapse, in which a single shared policy converges to the same control strategy across different morphologies.
ROOF proposes a reinforcement learning method with contact-supported geometric escape guidance that enables proprioception-only humanoid robots to first create clearance and then complete the rise under overhead obstacles. In simulation, the recovery success rate was 88.5% under plates that can only be lifted, compared with 18.9% without geometric guidance; it was 98.6% under free plates with randomized size, mass, and placement. After deployment on a Unitree G1, the policy successfully recovered in 95 of 150 attempts across seven real-world obstacle scenes, including furniture and multi-object arrangements not seen in training.
The study proposes Proof-Carrying Evolution (PCE), in which a language model provides both a controller and its certificate during search; only candidates that pass certificate verification can be returned, and the deployed artifact is the control law together with its certificate. The authors compare four language models with classical search and synthesis methods on speed-limited docking, maneuvering-target pursuit, and SO(3) attitude control. Certificates for linear problems are verified using exact rational arithmetic; on SO(3), floating-point verification is followed by exact re-verification of each reported champion. The best run of each method also receives a Lean 4 proof.
This paper proposes a separable catenary model for suspended-fabric Robotic Manipulation Surfaces (RMS) that can quickly compute the surface shape for given actuator heights, as well as the actuator heights needed to achieve a desired local slope. The closed-form inverse solution lets a simple PID controller position objects on the fabric surface, and the model relates fabric length, module span, and actuator stroke to the reachable workspace, yielding a lower bound on minimum fabric length and design charts.
The study proposes an online method for learning closed-loop dynamics under a frozen locomotion policy, using a contaminated Gaussian model to distinguish predictable from unpredictable responses and combining a Gaussian process with random Fourier features to build an online learner that reduces bias from heterogeneous observations. Combined with model predictive control (MPC), the method predicts floating-base motion more accurately; in simulation and hardware experiments, end-effector tracking error was 0.01 m, and a finite-sample error bound for the online learning is provided.
The research team proposes Movens-Zero, a scalable understanding-generation-execution architecture that supports real-time closed-loop humanoid motion generation under text, speech, music, and video inputs. Its accompanying dataset, Movens-5M, is built from Internet videos through clip generation, quality filtering, semantic annotation, motion recovery, and physical refinement with humanoid retargeting. It contains 15 million human motion sequences and 5 million physically executable humanoid motion sequences, totaling over 10,000 hours.
Why it matters: This technical report presents a complete path for building a humanoid motion dataset from Internet videos and training a unified understanding-generation-execution model, offering a reference for data scale and architecture design.
The paper uses higher-order kinematic derivatives of serial robots to propose an aspect-preserving real-time numerical inverse kinematics method, yielding the QuIK4 and QuIK5 solvers with fourth- and fifth-order convergence. The method also includes an aspect-preserving step limiter that extrapolates the Jacobian determinant along the step to predict and avoid singularities, as well as a certifier that uses Hermite interpolation to verify the aspect of solutions and a certified restart algorithm. The authors claim the failure rate is reduced by up to four orders of magnitude compared with existing methods, and they release a C++ library with ROS 2 integration.
This paper proposes a flexible sensing manipulation framework that directly connects the structural dynamics of flexible links with sensing, kinematics, and control. The method first identifies the dominant vibration modes of each flexible link according to response convergence and excitation bandwidth criteria. Under the condition of full-rank modal observability, m retained modes require only m independent slope measurements and one reference IMU, and the continuous link deformation is reconstructed accordingly for flexible sensing forward kinematics and a PDE-based controller.
The authors propose a decentralized motion planner based on model predictive path integral control that jointly computes the ego vessel's trajectory and predictions of other vessels' trajectories, encoding traffic rules as discontinuous costs without requiring smooth approximations. In maps taken from real sections of Amsterdam's canals, 780 randomized runs all reached their goals without collisions or deadlocks, outperforming an optimization-based planner designed for the same domain as well as a version of the proposed planner that does not account for interaction. Two scaled vessel prototypes completed head-on and crossing encounters without rule violations, planning and localizing onboard at 10 Hz in an outdoor environment.
The authors propose ∇Mimic, which uses exact gradients from a differentiable simulator to train whole-body motion-tracking policies for humanoid robots, reaching a score of 0.82 in 20 minutes and 0.874 in 1.8 hours on a single GPU, whereas the world-model method takes 1.2 hours to reach 0.82.
Why it matters: The paper presents a complete scheme for training humanoid policies with exact gradients under differentiable simulation and open-sources the engine and code, making it easy to compare the efficiency of first-order and zeroth-order methods.
The paper derives a complete analytical characterization of optimal collision-free trajectories for two robots moving from rest to rest under double-integrator dynamics and a squared-acceleration objective. It proves that these trajectories belong to eight contact families and gives the family determination for each combination of initial and final positions. The authors analyze the geometric properties of the optimal trajectories, provide a numerical computation algorithm, and discuss how these results can improve the performance of existing numerical trajectory optimizers.
The authors propose AC-SMPC, an actor-critic model predictive control framework with state constraints for safe navigation in constrained dynamic environments. The method represents the robot and obstacles as vertices and the workspace as half-spaces, locally linearizes the nonlinear dynamics, and convexifies hyperplane-based collision avoidance constraints into a differentiable MPC layer; the actor adjusts MPC objective parameters via closed-loop feedback, while the critic evaluates long-term task performance.
The paper proposes a bound-constrained kinematic calibration method to improve the positioning accuracy of a 6-DoF robotic needle-guidance manipulator. The method constrains a 19-parameter geometric identification model with physically plausible parameter ranges, treating the bounds as active optimization constraints to mitigate parameter coupling and overfitting. In simulation, end-effector RMSE dropped from 3.422 mm to 0.690 mm (a 79.8% improvement); in benchtop experiments, trajectory validation RMSE was 0.865 mm and coordinate grid validation RMSE was 0.610 mm.
DiffWBC proposes replacing deterministic policies with a one-step diffusion model for general humanoid motion tracking, requiring only one forward inference per 50 Hz control step, achieving a local tracking error of 20.1 mm on the BONES benchmark, and deployed on a physical Unitree G1.
The study proposes a joint-health-aware reinforcement learning framework that encodes actuator condition in a continuous joint-health vector and uses HealthNet to estimate each joint's health from recent observations and actions, allowing the policy to redistribute control effort across healthy and degraded joints.
The authors propose HIPC, a real-time humanoid iterative planning and control framework that couples a generative model with motion tracking into a unified planning-and-control paradigm. HIPC uses goal-conditioned motion planning and trains with a canonical motion representation to avoid gimbal-lock singularities under extreme poses, while mitigating out-of-distribution shifts caused by real-robot state feedback through two strategies: deviation-decayed recovery and history re-anchoring. Experiments show that HIPC achieves the best goal-reaching accuracy among existing methods and exhibits robustness to physical disturbances and the ability to recover after falls.
HoloBoxing proposes a humanoid boxing learning recipe that trains a policy on top of a pretrained whole-body controller, completing training from scratch in 7 minutes on a single consumer GPU, and has released the code and paper. The recipe separates reusable motor skills from task decision-making: a compact planner requests future body-point targets, while a frozen controller handles joint-level execution.
Based on a survey of more than 20 mainstream quadruped robot projects on GitHub, the article maps the open-source quadruped technology landscape across five schools: kinematic control, full-stack engineering, optimization algorithms, reinforcement learning, and hardware platforms.
Kinetix AI and KaiRobotics demonstrated a humanoid robot's ability to recover from falls in overhead-constrained spaces; the related research is titled "ROOF: Learning Humanoid Fall Recovery under Overhead Constraints." The content was presented as a Halloween-themed demo, emphasizing completing recovery motions in narrow spaces; specific robot model specifications, success rates, and experimental conditions were not disclosed.
Using Unitree H1, Isaac Sim 2024.1, IsaacLab, and PPO as its stack, the article walks through a complete sim-to-real pipeline: converting URDF to USD, scene setup, domain randomization configuration, PPO training, ONNX export, and ROS 2 deployment on the real robot.
Simplexity Robotics exhibited its i7 Pro and iX robots along with in-house arm and joint modules and other core hardware at the Sino-German Automotive Technology, Parts and Services zone of IZB 2026, held October 27-29 in Wolfsburg, Germany. The company presented embodied intelligence technology achievements centered on robot motion control, teleoperation and body development. Liu Minqiang, Director-General of the Investment Promotion Agency of the Ministry of Commerce, visited the booth for a tour and exchange.
Xuanji Dynamics' Hypertron series quadruped robots won the "High-Rise Fire Rescue Track" championship and first prize in the "Load-Bearing Climbing and Material Transport" category at the 2026 Guangzhou Safety Emergency Rescue Robot Competition. The Hypertron-SW01 is equipped with a self-developed high-torque joint motor module and a multimodal fusion perception system, features IP67 protection, and continued stable motion control, perception, and communication during a 3-minute continuous bake test at 120°C–130°C.
Humans should be more coordinated than robots. But that's not the case at SIGRobotics. @_robert_molina trained one of our robotic arms with reinforcement learning to bounce a ping-pong ball on a paddle continuously for up to 7 minutes. Watch this time-lapse video: it just keeps bouncing.
The study proposes REA (Routed Embodied Agent), a unified runtime that lets a reasoning-focused System 2, an embodied routing System 1, and reactive System 0 policies collaborate within the same framework, with the core rule that a running policy does not wait on the brain.
Taiying Technology Group has built an embodied AI robot mountain-climbing training ground at the foot of Mount Tai, replicating the 3.8-kilometer, more than 3,100-step terrain from Mount Tai's Hongmen Gate to Zhongtianmen, with 11 test tracks for humanoid robots and quadruped robot dogs preparing 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.
A team from the University of Naples Federico II and the University of Palermo used a spring-mass crawling model to analyze the movement of Pieris brassicae caterpillars on wood, steel, slate, and glass, finding that surface roughness is strongly correlated with crawling performance; glass is difficult to traverse because its microscopic roughness is insufficient for the prolegs to grip.
The author debugged and fine-tuned a motion controller based on the MPC/WBC framework for a self-developed quadruped robot, and used depth-camera images to build height maps. The robot can now climb stairs more reliably, step over obstacles up to 11 cm high (with 18 cm ground clearance), and handle soft ground with greater confidence. The controller is written in C++, is ROS2-compatible, and has been open-sourced in the voltdog/mors_quadruped repository on GitHub. Next, the author will focus on improving software reliability, including automated device connection tests and standing up from any posture.
Following the thread of "how to solve problems in different environments," the article reviews five categories of path planning paradigms in embodied intelligence: graph search (Dijkstra, A*, Hybrid A*), sampling methods (RRT, PRM), artificial potential field (APF), optimal control (MPC), and reinforcement learning and data-driven methods.
AUBO collaborative robots cover six major industrial scenarios—palletizing, assembly, grinding, spraying, inspection, and sorting—through force control, vision, and flexible deployment. The palletizing workstation supports one-click pallet pattern generation, 2-hour deployment, a maximum pallet height of 2.5 m, and a cycle rate of 9 boxes/min; the assembly station has a cycle time of 15 seconds and a fastening success rate of 99.8%; the newly released G40 adaptive collaborative robot has been used for grinding automotive metal structural parts and features IP67 protection and an explosion-proof design.
Mino's companion behavior application is implemented on the Jetson Orin Nano based on ROS 2, with three states: finding a person, following, and accompanying; the video is its first outdoor test. The application uses open-source modules such as depth estimation, YOLO-World person detection and tracking, pose feedback, odometry, and motion control, and a coding agent assembled the behavior logic within a weekend. The robot perceives the person by 3D position and velocity, smooths the trajectory with a Kalman filter, and continuously sends head pose and chassis velocity commands to custom hardware.
Yuequan Bionic's Bowen W-Bot 2.0 wheeled humanoid robot has been deployed in live-line distribution network operations and maintenance at State Grid in Hefei, independently completing the full live-line cable stripping process; the case was reported by People's Daily New Media's "Unimaginable Chinese Economy" program. The solution uses vehicle-mounted deployment on an insulated bucket truck, is equipped with the X-Hand M1 dexterous hand and 7-DOF humanoid arms, has a dual-arm payload of 20 kg and a whole-hand lifting payload of 50 kg, covers working heights of 0–2.2 meters, and has a chassis footprint of 0.238 m².
At the 6th Innovation Fair China-Europe Cross-Border Exchange Conference 2026, Zhujing exhibited its H1 humanoid robot "Xiao Zhu" and multi-scenario deployment solutions, with a live demonstration of writing the Chinese character "Fu," showcasing arm motion control and fine force-controlled manipulation. The company hopes to connect with European robot manufacturers, automation integrators, and research institutions to explore cooperation in industrial production, commercial services, and co-building high-quality datasets, and plans to work with ecosystem partners to build an overseas embodied intelligence ecosystem and drive the coordinated overseas expansion of its products and solutions.
At the 6th Innovation Fair China-Europe Cross-Border Exchange Conference 2026, Zhujing exhibited its H1 humanoid robot "Xiao Zhu" and multi-scenario deployment solutions, with a live demonstration of writing the Chinese character "Fu," showcasing arm motion control and fine force-controlled manipulation. The company hopes to connect with European robot manufacturers, automation integrators, and research institutions to explore cooperation in industrial production, commercial services, and co-building high-quality datasets, and plans to work with ecosystem partners to build an overseas embodied intelligence ecosystem and drive the coordinated overseas expansion of its products and solutions.