All updates
- ∇Mimic: Efficient Whole-Body Control for Humanoid Robots with Differentiable Simulation
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.
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.
arXiv cs.RO ·
- DMAT: Direct Action-Manifold Modeling for Efficient and Robust Bimanual Manipulation
The authors propose the Direct Manifold Action Transformer (DMAT), which reframes bimanual action generation as directly modeling clean, observation-conditioned joint-space action chunks from stochastic interpolants, enabling single-step inference without iterative denoising or trajectory straightening.
arXiv cs.RO ·
- RTPolicy: Real-time multi-view visuomotor policy for edge devices, inference at 32.9 Hz
RTPolicy is a multi-view visuomotor policy that uses view-aware feature fusion and visual compression to aggregate multi-camera information into a compact token representation, and uses a flow-matching action generator to reduce action-generation overhead. On the 50 tasks of the RoboTwin 2.0 single-task clean evaluation, it achieved an average success rate of 52.26%, higher than the strongest leaderboard baseline π0 at 46.4%.
arXiv cs.RO ·
- VLA policies still start executing without instructions: research on unintended task execution and start suppression
The study applied five types of instructions—empty, neutral, meaningless, infeasible, and wrong-task—to six model checkpoints in the LIBERO-Goal simulation, holding the initial state fixed and varying only the instruction, to examine whether the VLA policy starts, which task it completes, and what the instruction determines.
arXiv cs.RO ·
- Study provides complete analytical characterization of optimal collision-free motion for two robots
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.
arXiv cs.RO ·
- SCI: A Contact-Based Optimization Method for Collision-Free Motion Planning for a Million Robots
The authors propose the sequential contact insertion (SCI) algorithm, which organizes optimization around contact points where robots reach the minimum allowed separation and analytically reconstructs trajectories between contacts. In numerical experiments, the method can plan for one million robots within minutes, while thousands of robots take less than one second; on problems solvable by a fixed-grid sequential convex programming baseline, SCI achieves median speedups of several orders of magnitude with comparable solution accuracy.
arXiv cs.RO ·
- AC-SMPC: Differentiable Actor-Critic MPC with State Constraints for Safe and Fast Robot Navigation
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.
arXiv cs.RO ·
- SAFARI: Semantic- and Frontier-Driven Autonomous Risk Inspection Framework
SAFARI treats object detection as a native channel of a unified probabilistic map, letting geometric and semantic evidence jointly drive exploration and planning. The authors completed 11 missions on a quadruped platform across three field sites, achieving an average coverage of 98.7% of the accessible mission area, with 97.6% of mission time requiring no operator intervention, and detecting objects with characteristic dimensions as small as 15 cm. Ablation experiments show that object-size awareness and safety-aware path planning are critical for hazardous-object search.
arXiv cs.RO ·
- Physically Constrained Kinematic Calibration for High-Precision Robotic Needle Guidance
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.
arXiv cs.RO ·
- BioAgent: An embodied agent that autonomously acquires skills and expands capabilities for long-horizon laboratory protocols
BioAgent is an embodied agent for long-horizon laboratory protocols that acquires new skills by matching human-reviewed requirements with validated skills, selecting tool-based implementations for missing behaviors, and requesting human demonstrations as needed. On the second task, it reused 17 skills unchanged, modified 2 existing implementations, and added 5 new skill types; on two dual-arm laboratory tasks with 20 and 31 ordered skill invocations respectively, it achieved complete-task success rates of 66% and 58%, with each task run 50 times and autonomous recovery enabled.
arXiv cs.RO ·
- MolmoSpaces-CommonSense: A Commonsense Reasoning Benchmark for Evaluating Robotic Manipulation Policies
The authors propose the MolmoSpaces-CommonSense benchmark, which builds a reasoning taxonomy grounded in developmental cognitive science and examines commonsense reasoning about occlusion, containment, support relations, semantically correct grasp points, spatial affordances, and hidden objects.
arXiv cs.RO ·
- DiffWBC: One-step diffusion model enables steerable humanoid whole-body motion tracking
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.
arXiv cs.RO ·
- CalibSwitch: Observation-Driven Camera-LiDAR Extrinsic Recovery and Selective Maintenance Framework
CalibSwitch is an observation-driven camera-LiDAR extrinsic calibration framework that separates absolute extrinsic recovery from selective state maintenance.
arXiv cs.RO ·
- TaSP-KKT: Geometry-Aware Humanoid Motion Retargeting Based on Manifold ODEs
The study proposes TaSP-KKT, which models the continuous-time limit of OmniRetarget's per-frame KKT conditions as an ODE on the robot configuration manifold and implements motion retargeting through forward discretization. The method requires only one weighted tangent-cone projection and a closed-form manifold update per step, avoiding repeated inner SQP iterations. In numerical experiments, retargeting quality is comparable to OmniRetarget but slightly lower, while running speed is about twice as fast.
arXiv cs.RO ·
- Joint-Health-Aware Reinforcement Learning Framework for Robust Control Under Actuator Degradation
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.
arXiv cs.RO ·
- HandMimicX: Learning Object-Centric Dexterous Manipulation Policies from Egocentric Videos
HandMimicX proposes an object-centric framework for learning dexterous manipulation skills from egocentric human RGB-D videos. The method defines the reinforcement learning objective in terms of object motion, decoupling skills from specific hand trajectories, and imposes constraints on hand-object interaction states during grasping and manipulation to obtain more anatomically plausible hand configurations. Simulation and real-robot experiments show that the method learns efficiently from human videos and transfers skills to novel spatial configurations.
arXiv cs.RO ·
- HIPC: Humanoid Iterative Planning and Control via Goal-Conditioned Motion Closure
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.
arXiv cs.RO ·
- Hourglass: Cross-Embodiment Motion Planning by Coupling a Body-Agnostic World Model with Differentiable Ego-Dynamics
The paper proposes the Hourglass architecture, which decomposes cross-embodiment motion planning into a body-agnostic world model and each robot's differentiable ego-dynamics template, coupled only through a camera ego-motion interface. On seven simulated robots across five motion types, this generalist policy achieves success rates 3.8–10.6 percentage points higher than imitation learning and diffusion policy baselines on the same world model, matches each robot's specialized policy without dedicated training, and comes within 1.4–2.8 percentage points of a parameter-conditioned MPPI teacher.
arXiv cs.RO ·
- LEXP: Learning Robotic Manipulation from Physical Experience via Context-Space Adaptation
The authors propose LEXP, a framework for learning from physical experience through context-space adaptation that combines task-level planning with an experience module, reusing execution experience to generate contextual feedback while model weights and the action interface remain fixed. Across 12 real-world manipulation tasks, LEXP achieved 23/36 successes (63.9%), compared with 19/36 (52.8%) using memory alone and 2/36 (5.6%) for vanilla ChatGPT-6-Astra.
arXiv cs.RO ·
- TRACE: Target-State Risk-Gated Execution for Embodied Visual Tracking
The research team proposes TRACE, an execution-time target-state risk gate that intervenes selectively while keeping the learned policy unchanged. They first used PPO to update only the planning head of a 0.6B VLA tracker; success and following rates improved across all three EVT-Bench settings, but the collision rate rose in distracted tracking scenarios.
arXiv cs.RO ·
- CONSORT: Multi-Robot Cooperative Localization Planning with Uncertainty Constraints in GPS-Denied Environments
The authors propose CONSORT, a multi-robot planning method for cooperative localization in GPS-denied environments. By optimizing only over mean-state trajectories and propagating covariance only for terminal constraint evaluation, it turns the belief-space problem into a solvable form.
arXiv cs.RO ·
- TinyAstral: Transferring a self-evolving harness to small generalist manipulation agents
TinyAstral proposes a two-stage method: first freeze an open-weight teacher model and optimize its harness through recursive self-improvement (RSI), then distill a 4B student model to replace the teacher within the same harness. On RoboTwin, harness evolution raised the teacher's success rate from 33.4% to 63.6%, and the final 4B student reached 62.8%; in measured deployment episodes, the student used about one-fifth of the teacher's accelerator time.
arXiv cs.RO ·
- CC-HGCBF: Zero-Shot Safe Navigation for Heterogeneous Mobile Robot Teams
The authors propose the Capability-Conditioned Heterogeneous Graph Control Barrier Function (CC-HGCBF), which provides a unified safety representation for robots with differing kinematic and actuation capabilities in a shared second-order reference-point operational space. The method introduces the Radial Capability Field (RCF), which characterizes each robot's state-dependent finite-horizon reachable velocity set in a fixed dimension, enabling the shared graph barrier and decentralized policy to adapt to changes in state, kinematic constraints, and control limits without predefined robot identities.
arXiv cs.RO ·
- CA-BRF: A Capability-Aware Backup-Recoverability Filter for Heterogeneous Multi-Robot Navigation
The authors propose a Capability-Aware Backup-Recoverability Filter (CA-BRF) that constructs a backup-recoverable set in a shared Cartesian operational space and uses it to screen candidate controls, so that each robot retains a certified recovery maneuver consistent with its own kinematic and actuation constraints. The method selects admissible interpolations between candidate and backup actions and coordinates simultaneous interactions, thereby guaranteeing recursive feasibility and safety of the discrete-time closed-loop system even when the nominal safety-constrained optimization is infeasible. Simulations cover up to 30 heterogeneous robots, and experiments involve five distinct kinematic classes.
arXiv cs.RO ·
- Veloxity: Memory-Safe UAV Embedded Firmware Written in Rust
The authors present Veloxity, an open-source embedded UAV autopilot written entirely in the memory-safe Rust language, offering a minimal, modular foundation for autonomous drones. They compared it with equivalent C++ autopilots in simulation, on hardware testbenches, and in outdoor waypoint-tracking flight tests; the results showed functional parity with improved performance, indicating that memory-safe languages can be used for reliable, high-performance autonomous aerospace firmware.
arXiv cs.RO ·
- CIMI: Transferring Dexterous Manipulation Skills Through Contact Unfolding
The authors propose the Contact Imitation Interface (CIMI) framework, which uses contact geometry and fingertip RGB views to build a robot-free human demonstration interface and transfers skills to a robot hand through contact-alignment-first retargeting.
arXiv cs.RO ·
- Confidence-Modulated Safety Filtering for End-to-End Policies
The authors propose a filter-agnostic pipeline that modulates the conservativeness of a safety filter at runtime using policy confidence: inside the filter's operational design domain (ODD), confidence is traded for performance; beyond the ODD, low confidence yields additional safety margin. Preliminary F1TENTH simulations show that the modulated filter combines the strengths of permissive and conservative filters.
arXiv cs.RO ·
- From Teleoperation to Long-Horizon Dexterity: Evaluating Visuomotor Skills and Offline Proxies for Dexterous Manipulation
The research team used a five-fingered robotic hand to grasp and insert RJ45 connectors, evaluating 281 imitation-learning policies spanning seven policy families in approximately 1,700 real-robot trials. Controlled ablations showed that observation and action representations had a greater impact than conventional optimizer and sampling choices. For point-cloud policies, region-of-interest cropping, representation in the robot-base frame, and normalization within the point encoder each improved success by 70 to 100 percentage points. For the vision-language-action policy, switching to absolute action prediction improved insertion success by up to 70 percentage points.
Approximately 1,700 real-robot trials compared 281 imitation-learning policies, indicating that observation and action representations determine real-world success more than optimizer and sampling choices.
arXiv cs.RO ·
- CaPtain: A pipeline for post-training small open VLMs into code-as-policy robot agents
CaPtain proposes an open post-training pipeline that trains small open vision-language models (VLMs) into code-as-policy (CaP) robot agents, without relying on VLAs, frontier models, or privileged state at test time. The pipeline first performs supervised fine-tuning on synthetic trajectories generated by GPT-6 Astra, then applies multi-turn reinforcement learning, introducing a physics critic that reads simulator privileged state to assign credit to each request.
arXiv cs.RO ·
- Jiajie Electric: No robot products; carrying out digital transformation using robotic arms to replace manual labor
Jiajie Electric stated on an interactive platform that the company currently has no robot products and is carrying out a digital transformation in which robotic arms replace manual labor. Robotic arm applications have been implemented in three areas: stator core stacking, stator core welding, and rotor painting. This transformation is mainly used in the company's own production and manufacturing processes. R&D plans must be comprehensively determined based on factors such as market policies, industry development, and the company's strategic layout.
36Kr Embodied Intelligence ·
- Fixed 31 bugs in one week, only to find just 5 root cause categories
The 31 failures logged in the week before the competition, after deduplication, fall into 5 root cause categories: open-loop control, changes not taking effect, dead code, false feedback, and no attribution. Each root cause comes with real code and arithmetic reasoning, plus three actions you can follow directly.
Guyuehome ·
- Highlights from the Tencent Global Digital Ecosystem Summit
Highlights from the 2026 Tencent Global Digital Ecosystem Summit held in Shenzhen. Partners and technology leaders shared how AI is being integrated into daily work. #TencentGDES2026
腾讯 · X ·
- Taking Nine as a Sample: The Two-Wheeled EV Industry Has Reached the De-Storytelling Stage
Nine Company reported third-quarter revenue of 7.585 billion yuan, up 14.1% year on year, and net profit attributable to shareholders of 343 million yuan, but profit declined in the first three quarters and its valuation fell back to the industry midpoint. In the first half of the year, electric two-wheeler sales reached 2.844 million units, up 19% year on year, driven by entry-level lead-acid models trading price for volume, with gross profit per unit falling from about 676 yuan to 453 yuan; the lawn-mower robot business had revenue of 3.655 billion yuan, up 66% year on year, with a gross margin of 46.57%, but was weighed down by rising raw material prices and investment in overseas channels.
钛媒体 ·
- Trading money for time: embodied AI companies step into investing
Embodied AI companies are investing heavily in the robotics supply chain as industrial capital. Unitree Robotics and Sequoia China established the "Shanyu Innovation Program," and Zhiyuan Robotics appeared in 47 investment and financing events from August 2024 to October 2026, most of them directed upstream of its own complete machines. The article argues that three overlapping uncertainties—unconverged technology routes, an unformed supply chain, and unvalidated scenarios—are leading top companies to use capital to lock in supply chains, technology options, and industry positions in advance, as the industry moves from free competition toward an oligopoly rehearsal.
OFweek Robotics ·
- Meta partnering with a Chinese company to develop a humanoid robot? Fourier responds
Responding to reports that Meta has partnered with Fourier to build a humanoid robot prototype in about three months and is already developing a second version, Fourier founder and CEO Gu Jie said he is not yet able to comment. The rumored prototype, called Max internally at Meta, weighs about 48 kilograms and is intended for internal research and experimentation. Meta Vice President of Robotics Products Li-Chen Miller reportedly spent nearly two months at Fourier's Shanghai factory.
36Kr Embodied Intelligence ·
- How Alibaba Cloud E-HPC Supports Scientific Research Agents Through Supercomputing-Intelligence Integration: ToB Industry Observation
Alibaba Cloud is advancing supercomputing-intelligence integration through its E-HPC platform, using a unified computing network to simultaneously carry HPC, AI training and inference, and agent tasks, and has launched a four-layer E-HPC agent platform. Its RoboAgent connects the embodied intelligence chain from device-side collection, annotation, and training/inference to simulation deployment; HPDA Robotics includes the built-in APAK acceleration engine and claims that multimodal data processing cost-effectiveness can improve by 2.1 to 5 times. E-HPC has also formed three dedicated solutions for chip design, AI4S and life sciences, and embodied intelligence.
钛媒体 ·
- Rokae Robotics Wins 2026 Cailian Press ESG Zhiyuan Award Corporate Governance (G) Pioneer Enterprise Award
Rokae Robotics received the "Corporate Governance (G) Pioneer Enterprise Award" at the 2026 ESG Sustainable Development Value Ecosystem Forum hosted by Cailian Press on October 30, with the selection based on its governance structure, responsibility mechanisms, and decision-making system. In the first half of 2026, the company's revenue was 415 million yuan, up 136% year on year, and total robot sales exceeded 13,000 units; embodied intelligence robot revenue was 139 million yuan, accounting for 34% of total revenue. On September 29, the 10,000th humanoid force-controlled arm rolled off the line and was delivered, and the company has partnered with more than 80 industry partners.
珞石机器人 ·
- Robotics Concept Stocks Swing Lower; Unitree Robotics Falls Over 10%
Robotics concept stocks swung lower, with Unitree Robotics falling over 10%, Daye Co., Ltd., Zhengyu Industrial, and Tianqi Co., Ltd. hitting the limit down, and Shandong Mining Machinery and ASD falling over 7%.
36Kr Embodied Intelligence ·
- Fudan and Donghua team incubates Lingshu Zhifu, ionic-skin tactile sensors raise over 10 million yuan from Oriza and Huawei Hubble
Shanghai Lingshu Zhifu has completed consecutive seed and angel funding rounds, with participation from Oriza Holdings and Huawei Hubble, totaling over 10 million yuan across the two rounds; Linglu Capital served as exclusive financial advisor. Founded in April 2025, the company focuses on ionic-skin flexible tactile sensing materials and devices, with products including data-collection gloves and tactile sensors for dexterous hands. It has completed dual-85 testing and over 10 million repeated mechanical tests, is building a materials pilot line, and is sending samples to customers in multiple fields.
36Kr Embodied Intelligence ·
- Desktop robots are booming: embodied AI makers bet on a new track, is the endgame a physical Agent?
Desktop robots are becoming a new offshoot for embodied AI makers, focused on two functions: emotional companionship and productivity tools. The Letianpai desktop robot entered the market at a price of 1,799 yuan, the Lingtong Nian NIA-F01 offers companion interaction in a figurine form, and Elephant Robotics' myCobot series of small robotic arms requires secondary development. The article argues that consumer products still center on emotional companionship and have not yet truly delivered on boosting productivity, and that the future requires moving toward modularity and upgrading the ability to operate in the physical world.
OFweek Robotics ·