The paper proposes DriveDegrade, an image-degradation tolerance benchmark for camera-only end-to-end (E2E) driving. It injects 16 corruption types at 5 severity levels online in the image loader, uses a single operator to cover 15 policies, and evaluates open-loop planning on nuScenes and NAVSIM, with a CARLA closed-loop anchor.
Cyclops is a framework that synthesizes sparse Non-Repetitive Scanning LiDAR (NRS-LiDAR) intensity data into RGB video for all-weather perception tasks without a camera. Its method first uses a frozen pre-trained densification module to convert sparse intensity projections into dense representations, then learns a velocity field through Latent Bridge Matching (LBM) to transport intensity latents to the RGB distribution within a few ODE integration steps; it also injects prior-frame context via temporal attention to suppress inter-frame flickering, and models the velocity field as a policy optimized by a differentiable terminal reward.
FELT is a framework that uses existing paired visuo-tactile data to learn contact-aware representations from RGB observations, providing them to downstream policies through two interfaces: generated tactile images and contact-aware features. It combines a frozen visual encoder with separate query decoders for the left and right fingers, and the two branches exchange information through gated connections, with a convolutional readout producing spatial contact representations. On four contact-intensive real-robot tasks, both interfaces achieved higher success rates than vision-only policies, and directly integrating contact-aware features into policies raised success rates by 15-22.5 percentage points.
PinNet generates point cloud keypoints and their descriptors, and uses a plane-based geometric self-attention module to model spatial relationships between keypoints, thereby improving descriptor discriminability for loop closure detection and point cloud registration. The method was evaluated on datasets collected with various different LiDAR sensors; experimental results show strong place recognition performance, precise relative pose estimation, and successful single-shot localization in different environments.
HyVGGT-VO proposes a hybrid dense visual odometry framework that combines the high-frequency pose output of traditional sparse VO with the dense reconstruction capability of the feed-forward model VGGT, and is claimed to be the first work to tightly couple a traditional VO frontend with VGGT.
PanoAir proposes a panoramic visual-inertial SLAM (VI-SLAM) framework and builds a real-world panoramic dataset covering different illumination, altitudes, trajectory lengths, and motion dynamics. The method uses panoramic feature extraction and panoramic loop closure to strengthen feature constraints and global consistency. On its own dataset and public benchmarks, it completed all 17 sequences without tracking failure and reduced ATE by up to 47.1% compared with the best baseline. Deployment on an embedded platform verified computational efficiency comparable to the PC implementation. The source code and dataset are publicly available.
The StopWatch paper diagnoses goal grounding as the main failure cause of current object-goal navigation policies based on vision-language models (VLMs): these policies explore well but often fail to notice that the goal is already in view or to estimate its distance.
The paper releases TriSAR-3D, a trimodal person detection dataset collected with a 128-channel LiDAR, an RGB camera, and a thermal camera in indoor and outdoor settings at different times of day, covering diverse poses including lying down; it also proposes a fusion method, PaintFusion.
TransNav proposes a Transformer-based end-to-end navigation method for quadrotors that achieves obstacle avoidance in environments with mixed static and dynamic obstacles. The method uses a patch-based point-cloud encoding scheme, integrates perception and navigation into a single architecture, and directly predicts a finite-horizon action sequence and executes the first action. It also introduces a lightweight point-mass dynamics model that optimizes trajectory-level costs through numerical gradient propagation, enabling unsupervised training without expert demonstrations.
The paper integrates FOUND-IT into a mobile manipulation robot platform, validating that foundation-model-based 3D scene graphs can support object change detection, multiple mapping sessions, navigable path determination, and object grasping in scene rearrangement tasks. Based on rigorous real-robot experiments, the authors provide an analysis of the effectiveness and limitations of a monocular, foundation-model-driven 3D scene graph construction system. The code is open-sourced in the MIT-SPARK/FOUND-IT repository on GitHub.
F-LIVO proposes a filter-based LiDAR-inertial-visual odometry (LIVO) framework that couples the visual-inertial and LiDAR-inertial subsystems through feature-level point-to-plane constraints between visual points and LiDAR planes.
FLoTS-SLAM proposes deriving motion uncertainty from the residual between optical flow and camera-motion-induced flow, for monocular RGB SLAM in dynamic scenes. The system first estimates camera pose using high-confidence pixels, then reprojects dynamic pixels to jointly optimize pose and depth, reducing the impact of object motion on pose and depth updates. On DAVIS, the average IoU of dynamic masks improves by 35% over advanced baselines; on Bonn and TUM, average ATE is reduced by 59% and 19%, respectively, relative to the SLAM backbone used. The source code will be made public.
The paper proposes ActiveAlign, a data adaptation framework that addresses the mismatch between the initial observation state produced by active camera motion and the manipulation policy's training distribution (first-frame distribution sensitivity): on LIBERO, directly prepending camera repositioning to pi_0.5 reduces the success rate from 46.3% to 11.5%.
SkyAgent proposes Revisable Target Memory (RTM) to address the problem in continuous aerial search where, after evidence becomes contradictory, old candidates still trigger approach or verification actions (which the authors call false candidate persistence, FCP). RTM associates evidence with persistent candidates, suppresses repeated counting of correlated views, and lets candidates be revised among three states: Tentative, Confirmed, and Rejected.
The paper proposes a gradient descent bio-inspired magnetic navigation (BIMN) method based on local gradient estimation with a Rao-Blackwellized particle filter (RBPF) for navigation in GPS-denied environments. The method does not rely on complex, bulky, and costly magnetic maps; gradient descent naturally draws the system toward the destination, enabling magnetic navigation without magnetic maps as a simple and memory-efficient approach.
PASS is a trustworthy precision approach simulation system for fixed-wing aircraft CAT III integrity algorithm development. Within a unified spatiotemporal geometric framework, it integrates configurable airport environments, prescribed trajectories, inertial simulation, calibrated stereo images, and continuous runway-relative ground truth, supporting navigation and integrity research under controllable sensor errors and environmental variations.
The study proposes GR-VIP, a goal-conditioned visuomotor policy that uses a visual encoder pretrained for relative camera pose estimation, conditioning action prediction on dense relational features before pose regression to preserve fine-grained cross-view information for learning corrective actions from demonstrations. In real-world docking tasks, GR-VIP achieves high docking success rates and millimeter-level terminal accuracy in both the training environment and unseen environments, with transfer requiring no additional training. Supplementary experiments show the policy can align with non-docking targets and complete docking under the low-light conditions tested.
Ego2Contact learns visual-tactile priors from human interactions and transfers them to robot hands. Its predictor fuses temporal scene context with hand-local visual features to infer bimanual tactile fields from RGB video, and anatomical region correspondences align varying robot tactile layouts with the predictor's canonical grid.
A CoRL 2026 paper proposes SpatialBEV, which queries bird's-eye-view (BEV) features around the predicted state at each step and conditions the next prediction on current geometric and semantic cues, in order to capture the complex interactions between terrain and vehicle in off-road scenes.
PASS-VLN is a passability-aware benchmark and model for local-command vision-and-language navigation (VLN). It holds the scene, instruction, start pose, and reference route fixed while varying only the robot body width, requiring the robot to reach the target when the passage is feasible and to stop early when it is blocked.
The arXiv paper "Spatial Memory by Agents, for Agents" investigates letting coding agents automatically design robot spatial memory: under a fixed development budget, a Designer Agent writes and iteratively revises memory construction and retrieval code, with feedback from a fixed Query Agent; the resulting implementations are frozen and then used to build memory from held-out observation histories for evaluation.
The arXiv paper Micro-Localization demonstrates high-precision global localization using microscopic optical observation of natural ground texture, achieving localization accuracy on the order of tens of micrometers. The authors established the relationship between optical magnification and image-space localization uncertainty on the one hand and physical localization accuracy on the other, and addressed the calibration, camera modeling, and focusing challenges introduced by high-magnification imaging.
The paper proposes RTISSLAM-3D, a 6-DoF SLAM system for a wheeled robot equipped with an in-air 3D imaging sonar, an IMU, and wheel encoders. The system uses a spherical CFAR detector to reduce sonar images to sparse 3D detections, which are used for 3D acoustic odometry that models the sonar beam shape, for place recognition with polar submap descriptors, and for loop-closure verification via 3D registration, all combined into a 6-DoF pose graph with attitude factors.
SpatialScout is a new benchmark for evaluating vision-language model (VLM) agents' ability to actively build object-level 3D maps. It contains 80 diverse indoor scenes and 3,140 target objects. The task requires agents to control a camera to explore scenes and estimate object centers in a shared coordinate frame, coordinating viewpoint selection, cross-view association, geometric estimation, and map revision within a limited interaction budget.
Seerbotics demonstrated its self-developed VLA model driving a robotic arm to complete the seven-move sequence of a three-disc Tower of Hanoi. When an observer slides the base to change its overall position or rearranges the discs into an arbitrary intermediate state, the robotic arm can still re-identify the current configuration, determine the next move, and continue completing the remaining task. Trained on only 60 demonstration trajectories, the model can generalize across changes in object position, color, and shape. The Seerbotics robot "brain" consists of three layers—model, cerebellum, and nerve—responsible respectively for scene understanding and decision-making, motion planning and execution, and a unified interface connecting sensors and joint actuators.
Hyundai Motor and Kia will hold the seventh E-FOREST TECH DAY 2026 from October 19 to 21 at the Uiwang Research Institute, showcasing about 200 new technologies covering software-defined factories (SDF), automation, energy, and future mobility.
Shining 3D has launched the FreeScan Combo Wireless series of wireless metrology-grade dual light source handheld 3D scanners, including the FreeScan Combo+ Wireless and FreeScan Combo Wireless. The body has a net weight of 550g, features a built-in wireless transmission module and battery, comes standard with 3 swappable batteries, and supports an external power bank.
The paper proposes the HumanoidPose framework, which treats vision as proprioception and recovers a humanoid robot's full-body joint configuration in world coordinates from video alone. The method combines pretrained human pose features with robot kinematics in a unified model and retargets human motions to sixteen different humanoid embodiments, enabling scalable learning entirely from synthetic data; it also uses unlabeled online videos through self-training to develop a 2D robot pose detector to improve real-world generalization.
A 4-DOF 3D-printed DIY robotic arm has for the first time successfully recognized an opponent's moves via camera and autonomously completed a game of Nine Men's Morris. Joints J1-J3 use POM gears from Tarot 450 helicopter main gears paired with NEMA 14 stepper motors, J4 is driven by an RC servo, and the gripper is made of flexible TPU; the controller is a BTT SKR Pico 3D printer board with TMC2209 drivers supporting sensorless homing.
A project plan for using drones equipped with RK3588 for ship detection and on-site verification in shipyards, covering image capture, model training, on-board inference, and evidence preservation. The plan includes five scenario images showing detection boxes and explains occlusion handling and the deployment verification process.
The "3D Vision from Beginner to Mastery" Knowledge Planet has been running for nearly 9 years and now has more than 6,400 3D vision practitioners. It includes nearly 20 core video course tracks, covering industrial 3D vision, SLAM, 3DGS, autonomous driving, drones, embodied intelligence, and other directions.
A team at the Hong Kong University of Science and Technology developed the embodied robotic arm EmArm, embedding a soft optical tactile skin into a rigid skeleton to achieve whole-arm tactile sensing and adaptive manipulation, with tactile sensing fidelity reaching 97.9% under five environmental disturbances. Its tactile skin covers over 71,000 square millimeters with a spatial resolution of 0.085 millimeters, can recognize five touch behaviors and provide emotional feedback, with recognition accuracy up to 99.8%. The research was published in Nature Sensors.
The REEMAN dual-laser Fly Boat factory delivery robot (Fly Boat AMR, ABOT11B) is deployed at an electronics factory in Zhuhai, transporting about 200 kg of PCB materials and bottled water per trip, with a rated payload of 300 kg, a loaded delivery route of about 600 m plus an empty return of about 100 m, and the customer reports about 70–90 trips per day.
On October 14, Hesai launched the Kosmo 3D world data engine, including the Kosmo One handheld AI spatial camera and the Kosmo World cloud platform, aiming to make spatial modeling as easy as taking a photo.
Scale AI, in partnership with Elorian, released a new benchmark called Humanity's Sixth Sense (HSS) to test intuitive visual reasoning in multimodal models: 20 human participants scored 93.1%, while top-ranked GPT-6-astra reached only 53.6% even at maximum reasoning effort, and the median across 25 models was 30.9%.
Integrated Systems Design (ISD) last week released its Palletizing Module, which can perform outbound palletizing and inbound depalletizing at the same station, switching modes without reconfiguring hardware.
In machine vision, CNNs and Transformers divide work according to real-time performance and complexity: CNNs handle lane recognition, close-range obstacle detection, traffic sign classification, and workpiece positioning and defect detection for industrial robots, while Transformers handle bird's-eye-view scene reconstruction, multi-sensor fusion, trajectory prediction, and drivable-area segmentation.
A solo developer spent six months turning Visual Recovery, which was still just an idea in April, into a general-purpose real-time 4K 60FPS NVIDIA CUDA underwater image enhancement software, with algorithms designed specifically for real-time processing in turbid water. In this test, the GPU took about 5-10 milliseconds to process each frame. The project is still experimental, and the developer is seeking feedback from ROV operators and underwater computer vision practitioners.
EvergreenSLAM has released the v0.3-preview preview version. The author states that this is still a preview, not a product, and that which workflows have run on the robot should be checked in the Validation and limitations section of the README. Compared with v0.2-preview, the agent layer adds egs mcp, with the CLI provided as MCP tools over stdio and every command supporting --json, and adds a pose timeline GET /pose, relocalization results, and recovery guidance for agents.
A two-stage joint optimization calibration method for LiDAR-camera extrinsics assisted by monocular depth is proposed, replacing traditional grayscale-intensity mutual information with depth constraints to improve calibration accuracy and robustness under complex lighting and material conditions. In the coarse calibration stage, a composite calibration board combining ArUco markers and a circular-hole array is used, and an SVD-optimized ICP algorithm solves for initial extrinsic values analytically without requiring preset initial values; in the fine calibration stage, the DepthPro monocular depth estimation network is introduced to construct a joint optimization objective function combining geometric constraints and normalized mutual information distance (NID).