On October 26, Agibot launched the safety-certified heavy-duty embodied robot Genie G2 Max (AGIBOT G2 Max). Official figures are 20 kg per arm and 40 kg for both arms (50 kg peak). Its 7-DOF force-controlled dual arms achieve repeat positioning accuracy of ≤±0.5 mm, with vertical operation covering 0.1 m to 2.5 m.
Shenmou Intelligent's "Dragon Claw" aerial-operations drone has won State Grid Power's "High-Stability UAV for Aerial Operations" project, using a drone-plus-robotic-arm-plus-dexterous-hand configuration to address foreign-object hazards on overhead transmission lines. The system identifies foreign objects such as bird nests through inspection, then uses the robotic arm and dexterous hand to carry out integrated detection, handling, and verification. Shenmou says its self-developed technologies include the drone and robotic arm bodies, a multi-redundant flight control system, target grasping control, dynamic visual guidance, and piezoelectric six-axis force sensors.
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.
Agibot launched the safety-standard heavy-duty embodied robot Genie G2 Max on October 26, targeting depalletizing and palletizing, material transfer, and production-line feeding. It has a 20 kg single-arm payload, a 40 kg dual-arm rated payload, a 50 kg peak payload, and repeat positioning accuracy within ±0.5 mm. The robot uses 7-DOF force-controlled dual arms, has a vertical working range of 0.1 to 2.5 meters, features dual-battery hot swapping and autonomous recharging, and is compatible with the Genie Studio 2.0 embodied work platform. Agibot disclosed that in e-commerce warehousing scenarios its overall task success rate exceeds 99%, with 80 boxes depalletized and 75 boxes palletized per hour.
Atomrobot has launched a flexible parallel loading and unloading workstation for small new-energy lithium-battery parts, 3C electronic components, and small daily-chemical products, performing visual recognition and positioning, high-speed picking, and tray placement. The equipment requires no modification to the production line foundation, can start production with AC220V, and its graphical interface lets ordinary operators complete debugging; it also supports rapid changeover between products. The workstation has already served customers including CATL, BYD, Foxconn, Luxshare, and Sunwoda.
CASBOT completed an A++ funding round of several hundred million yuan, jointly invested by Zhilai Investment, Yonghua Investment, Maotian Investment, Yuanhe Chenkun, Guoke Kezhi, Luning Investment and others. The funds will be used for product mass production and delivery, R&D iteration of embodied brain technology, and market expansion. The company's self-developed CASBOT PA perception-to-action model has been validated at actual workstations in a shoe factory in Putian, achieving a cycle time of 7 seconds per piece for the outsole stamping process and 18 seconds per pair for sole pressing. In 2026, its total orders for embodied shoe-making robots exceeded 400 million yuan, with equipment production scheduling reaching the scale of several thousand units.
Why it matters: The funding announcement also disclosed shoe-making workstation cycle times and order scale, which can be used to assess the industrial deployment progress of embodied manipulation models.
Shenmou Intelligence's Dragon Claw aerial work drone has won State Grid Power's "High-Stability UAV for Aerial Operations" project, using a drone-plus-robotic-arm-plus-dexterous-hand configuration to clear foreign-object hazards on overhead transmission lines. The system identifies foreign objects such as bird nests through inspection, then uses the robotic arm and dexterous hand to carry out removal and verification, forming an integrated discover-handle-verify workflow. Shenmou says it developed the flight control system, robotic arm and dexterous-hand grasping control, dynamic visual guidance, and piezoelectric six-axis force sensors in-house, and cites research estimating annual procurement demand of about 150 million to 200 million yuan per province.
Afferent-Omni is an open-source visuo-tactile command and control (C2) framework for embodied intelligence and dexterous manipulation, offering real-time visuo-tactile perception, closed-loop manipulation, and embodied AI evaluation capabilities. Its FastAPI backend includes 9 modular routers and 15 REST endpoints, and achieves closed-loop operation via a 50 Hz WebSocket stream; it integrates a neuromorphic reflex arc with under 15 ms latency, multi-finger topology, and acoustic friction profiles.
Topstar's humanoid dual-arm robot AOI glass-wiping workstation uses ±0.05mm repeat positioning accuracy, combined with visual positioning and flexible force control, to achieve targeted wiping without damaging glass in optical workshops. The workstation integrates inline AOI inspection and wiping in a single unit, replacing the traditional two dedicated machines and two workstations, and relies on a high-resolution industrial camera and an edge AI computing unit to perform defect detection and classification positioning. It has now been validated in the workshops of a leading 3C customer and an optical customer; ordinary production line workers can quickly learn to operate it, helping customers improve overall workshop yield.
The authors propose ContactBench 1.0, which uses 36 reference tasks and more than 40 object assets to break contact-rich manipulation down into eight essential skills; the assets can be reproduced on consumer-grade 3D printers and each comes with a matching simulation model. They evaluate recent vision-only and visual-tactile policy learning methods on six benchmark tasks, and the results show that vision-only policies struggle to perceive or respond to contact events, while visual-tactile policies do not bring noticeable improvement. The benchmark will be open-sourced, providing physical and simulated assets and controlled task variations for evaluating the generalization ability of robotic systems.
The research team proposes DASH, a da Vinci adapter for serial-link and humanoid robots that mounts da Vinci Classic and Xi instruments onto general-purpose robots without modifying the instruments themselves. The two adapter types are compatible with robotic platforms of various load capacities, engage the instruments' native latches, and wirelessly identify inserted instruments to load the corresponding kinematics and coupling matrices. Teleoperation experiments validated the approach and its role in lowering the barrier to surgical robotics research.
The authors propose RMMBench, a mobile manipulation benchmark that requires robots to understand language instructions and execute long-horizon tasks in continuous space, with a "navigation-manipulation" task suite covering 70 typical task scenarios. Results show that leading vision-language models (VLMs) still face notable challenges in spatial localization for mobile manipulation tasks, and their spatial perception in long-horizon interactions needs strengthening. The work has been accepted to CoRL 2026.
PartialBiGrasp is a dual-arm grasp generation framework that operates directly on partial point cloud observations. It implicitly learns geometric features through a convolutional occupancy network to reason about graspability, collision-free contact regions, and object thickness. Based on this, the method generates force-closure-compliant grasp pairs and uses sampling-based optimization to correct for ambiguity caused by incomplete geometry. The authors evaluated the method using analytical force-closure metrics, large-scale simulation experiments, and noisy partial point clouds on a real robot.
UCA-Flow proposes a unified condition-action modeling framework that encodes observation conditions, timestep conditions, interval conditions, and action tokens into a single sequence, which a unified Transformer learns jointly to enable accurate one-step action generation.
RAPAC-DP proposes a response-aligned pending-action compensation framework that enables diffusion and flow-based action generators to still output full action chunks aligned with the handover moment under cloud inference latency. The method uses a lightweight pending-action encoder and LoRA adapters to learn the compensation while keeping the base policy frozen, and constructs delay-conditioned training samples from delay-free demonstrations, requiring neither an explicit dynamics model nor additional delay data.
The authors propose ENAP (Emergent Neural Automaton Policy), a neuro-symbolic imitation learning framework that uses adaptive clustering and an extension of the L* algorithm to learn a Mealy state machine from demonstrations as an interpretable high-level planner, which then guides a low-level reactive residual network to learn continuous control via behavior cloning.
DexPort is a hand-held data collection interface for high-DoF robot hands that uses the same hand hardware for both collection and deployment, recording joint target commands and measured joint states. The operator moves the hand-held interface to specify wrist motion and coordinates finger motion by selecting or adjusting skills, without needing to control each joint individually.
The study proposes an OCT-guided data-driven framework that learns soft tissue shape control from user demonstrations and executes it with closed-loop OCT feedback. In 12 configurations and 576 phantom trials, the Explicit policy using residual-conditioned voxel point clouds achieved Shape RMSE of 0.534 mm on the depth task and 0.443 mm on the directional task; full-volume Diffusion with residual conditioning reduced directional error from 0.748 mm to 0.555 mm.
The authors propose Backtrack, a lightweight failure detection and recovery mechanism for frozen VLAs that requires no recovery demonstrations, policy retraining, or additional inference models. It uses a failure prediction head that adds only 0.48-0.88% to the parameter count to read the VLA's internal representations; when the predicted failure probability exceeds a threshold, it pauses the policy, backtracks the robot arm's recent motion, and hands control back to the original policy for a retry.
The study proposes an anatomy-anchored imitation learning framework that combines anatomy-guided probe placement with learned ultrasound navigation. Across 16 unseen participants, task success rates for standard-view acquisition of the kidney, gallbladder, and liver and for radial artery scanning were 96%, 94%, 88%, and 94%, respectively.
CoCarry introduces a dataset and a contact-support-structured retargeting method for dual-humanoid cooperative carrying. The dataset contains 220 synchronized dual-person sequences spanning 21 task schemes, 5 shared-object geometries, and 4 terrain types, totaling about 61 minutes of motion data.
DynaSlots is an action-label-free 3D visual pretraining framework that decomposes each point cloud frame into a small set of temporally consistent slots and models dynamics at the slot level rather than the scene level. A temporal slot propagation module keeps slot assignments consistent across frames; an inverse dynamics model extracts a latent action for each slot from feature differences; and a diffusion-based forward dynamics model, paired with a self-supervised interaction head, predicts each slot's future state and learns pairwise proximity and motion relationships.
MAST-VLA proposes a dual-conditioning framework that introduces skill-conditioned LoRA Mixture-of-Experts adapters into the LLM backbone and converts predicted affordance heatmaps into a patch-level trajectory-affordance spatial prior for the action head. In the CALVIN ABC→D long-horizon multi-skill cross-scene setting, it achieves an average completion length of 4.65 over five-subtask sequences, the best performance in that setting. Real-robot multi-skill sequence experiments further validate its deployability under real perception and control noise. The paper has been accepted to CoRL 2026.
The research team deployed a VLA model on a Unitree Go2 quadruped equipped with an AgileX Piper robotic arm, achieving closed-loop mobile manipulation, and reported choices such as model variant, action horizon, and demonstration conventions that made real-robot deployment feasible. For the pick-and-place task, expanding demonstration coverage reduced held-out end-effector error by 30%, and adding secondary skills such as pushing introduced new behavior modes without degrading baseline performance. Closed-loop real-robot experiments showed that dataset composition affects task success.
RoboCode learns short-horizon motion patterns that can be reused across tasks through a shared soft-gated primitive action interface, representing action chunks as soft compositions of multiple learned primitives rather than a single discrete code. The interface is learned on action chunks pooled across multiple tasks and then frozen, after which task-specific state encoders are trained to map into the shared latent space. On five ManiSkill tasks, it achieves an average closed-loop success rate of 75.6%, above ACT's 68.8% and Diffusion Policy's 41.6%, with an average action chunk reconstruction MSE of 0.0187.
The paper formulates robot-assisted dressing as a geometry-aware insertion problem and proposes a reinforcement learning framework combining a staged curriculum with a finger-wise insertion soft gate. In post-training evaluation on the deformable bracelet task, the full method achieved a success rate of 84.7±1.2%, with the insertion gate contributing the most and the highest score when combined with the curriculum. On a dummy hand, the same policy succeeded in 15 of 20 real-robot trials.
DiscoDemo uses a latent-conditioned reinforcement learning generator with privileged information, together with a parallel reverse curriculum and a success-gated diversity objective, to produce multiple successful closed-loop behaviors; after safety filtering, these are exported as a static dataset for downstream visuomotor imitation learning. On four manipulation tasks, the average downstream success rate in simulation rose from 40% for the strongest baseline to 62%, and on real robots from 39% to 78%. The authors believe that the diversity of successful solutions can improve imitation learning for pretrained flow-matching VLAs and complements corrective data collection.
EgoForesight is a vision-language-action framework that jointly leverages robot demonstration trajectories and noisy egocentric videos for policy learning, reducing the impact of unreliable action labels through reliability-conditioned corruption and reliability-weighted regression, and supplementing supervision from demonstrated outcomes via privileged-future representation distillation.
FORCA is a coding agent that takes visual observations and wrist force/torque measurements as input, writes Python programs for contact probing and force-responsive motion, and between executions uses force records together with visual outcomes to diagnose failures and revise motion and force-feedback control logic.
The authors propose the SEVI framework, which uses an evidential safety critic to simultaneously estimate future constraint-violation risk and epistemic uncertainty in a single forward pass, judging before an action is executed whether it is unsafe or insufficiently supported by evidence, and overriding that action with an offline-trained recovery policy.
PrefO2O proposes an offline-to-online latent preference transfer framework that enables zero-shot preference-based reinforcement learning (PbRL) on target tasks without any human feedback. The method learns a task-invariant latent preference space from a small amount of offline preference data on source tasks, disentangles transferable preference structures from task-specific variations, and uses confidence-weighted optimal transport alignment to generate pseudo-preference supervision for online reward adaptation.
Longhe Intelligent has launched the 'Speed' heavy-duty smart loading robot for high-tempo shipping scenarios. It autonomously completes the entire process from grasping and palletizing to full truck formation without manual intervention, cutting loading time per truck to 3.5 minutes, a 600% efficiency gain over traditional loading methods. The equipment can operate around the clock and adapts to domestic road transport vehicle types such as container trucks, flatbed trucks, curtain-side trucks, low-sided trucks, and high-sided trucks without modification, ready to use on installation. The overall technology behind this full-truck loading solution fills a domestic gap and can replace imports.
At the 26th China International Industry Fair, CAS Fifth Century used three Jiling-L1 wheeled dual-arm robots to relay a continuous material flow of depalletizing, pouring, receiving, and palletizing, with no manual intervention throughout. At the same event, it launched the column-type embodied intelligent robot Jiling-S1, featuring 19 degrees of freedom and 360° waist rotation, and introduced its next-generation embodied brain Jiyuan Lingnao, which relies on its self-developed FAM series embodied large models and motion optical flow technology to explore cross-embodiment transfer and few-shot learning.
Heli Yufeng has designated semiconductors as a strategic focus industry and, based on "stereoscopic storage + AGV handling and delivery," is building a full-chain logistics coordination system covering raw material storage, production line delivery, material picking, finished goods storage, and intelligent truck loading. The company has delivered an intelligent warehousing and conveying system integrating automatic depalletizing, inbound storage, picking, and outbound storage to a semiconductor chip manufacturing enterprise, and has provided an automatic packaging + intelligent warehousing solution for a semiconductor materials enterprise, achieving automated linkage between packaging and warehousing for anti-static protection and full-batch traceability.
Geek+'s general-purpose humanoid robot Gino 1 will appear at CeMAT Asia 2026 from November 3 to 6, 2026, at Booth A3-2, Hall N1, Shanghai New International Expo Centre. The on-site demonstration will showcase multi-category picking and cross-task transitions, combining multimodal perception and force-control feedback to determine grasp positions and strategies in real time, and it is equipped with the self-developed Gravity embodied intelligence model; under full-body parallel control, peak throughput can reach 300UPH.
Topstar's humanoid dual-arm robot AOI glass wiping workstation has completed validation in the workshops of a leading 3C customer and an optical customer, integrating the functions of two stations—inline AOI inspection and wiping—into a single machine. The workstation has a repeat positioning accuracy of ±0.05mm; it locks onto stain positions via vision and performs wiping with flexible force control, relying on a high-resolution industrial camera and an edge AI computing unit to complete the full chain from image acquisition to defect classification and localization. Ordinary production line workers can quickly get up to speed, reducing training costs and human error and helping improve overall customer workshop yield.
Rollback proposes a policy improvement paradigm that turns an action-conditioned world model from a one-way simulator into a backtrackable counterfactual engine: it returns to key decision points, branches to try different actions, and compares the outcomes, thereby distilling counterfactual results into action-level supervision without learning a separate value model.
Why it matters: The world model is used as a backtrackable counterfactual engine to provide fine-grained action-level supervision for policy improvement, and the improvement margins on three real-robot tasks are reported.
Honglu Steel Structure announced on October 23 that it has received another order from Henan Mine for its "nine-axis robotic welding workstation," with the customer adding intelligent welding equipment on top of the first batch of devices. Earlier in September, Honglu Steel Structure also successively won orders for cantilever dual-machine shared-rail, 8-axis and 9-axis cantilever welding workstations, and on October 22 reached a batch procurement cooperation with a steel structure enterprise covering 8 eight-axis inverted cantilever workstations and 10 upright floor-rail workstations.
Teradyne Robotics and Sereact have formed a partnership under which Sereact's Cortex physical AI platform will drive Universal Robots robot arms for unstructured tasks such as returns handling and picking; the two parties have already delivered dual-arm systems made up of hundreds of UR robot arms to customer sites in Europe.
ETH Zurich's ANYmal quadruped robot solved a 10-pound giant Rubik's Cube using only its feet, three times, taking about 8 to 9 minutes each time. The demonstration combines computer vision, a traditional Rubik's Cube solving algorithm, and reinforcement learning to teach the robot to balance, apply force, and correct errors. This "pedipulation" (manipulation with feet) could enable future robots to open doors, turn valves, or move obstacles without hands.