#LiDAR (71)

Waterfall Pagination

A Humanoid That Chooses Contact Points to Cross Terrain

A Humanoid That Chooses Contact Points to Cross Terrain

@ZappyZappy7 · 151 views · 2026-10-25

HumanoidLocomotion

Hesai Kosmo: LiDAR + 8K Capture for Cloud 3D Reconstruction

Hesai Kosmo: LiDAR + 8K Capture for Cloud 3D Reconstruction

@tokufxug · 125 views · 2026-10-25

3D reconstruction3D Gaussian Splatting

Catching LiDAR Ghost Cars by Their Missing Shadow

Catching LiDAR Ghost Cars by Their Missing Shadow

@AlloyRobotics · 126 views · 2026-10-24

LiDARAdversarial Attack

NVIDIA Cosmos 3 for Driving: 11 Synced Camera Views and LiDAR

NVIDIA Cosmos 3 for Driving: 11 Synced Camera Views and LiDAR

@tokufxug · 154 views · 2026-10-25

ACP: Humanoid Ladder and Platform Climbing via Attention-based Contacts

ACP: Humanoid Ladder and Platform Climbing via Attention-based Contacts

@CyberRobooo · 171 views · 2026-10-23

HumanoidLegged Locomotion

Into the LidarVerse: A Real-Time Unified LiDAR World Model Across Tasks and Sensors

Into the LidarVerse: A Real-Time Unified LiDAR World Model Across Tasks and Sensors

We present LidarVerse, a real-time unified LiDAR world model for sensor simulation across tasks and sensors. Existing LiDAR simulators largely follow one of two complementary paradigms: reconstruction-based methods preserve the geometry of a captured scene but require lengthy per-scene optimization, whereas generative methods amortize learning across data but lack the consistency needed for long-horizon rollouts. LidarVerse brings both paradigms into a single, unified conditional generative framework. A universal tokenizer maps LiDAR range sequences into a shared latent space, where a bidirectional teacher learns a common interface for conditioning on a LiDAR prefix, an HD-Map, a persistent spatial cache, and optionally, text. Varying the available scene evidence turns the same model into a generator of new worlds, a point-cloud forecaster, or a generalizable resimulator of observed scenes without per-scene optimization. We distill the teacher into a four-step autoregressive causal student and maintain both a rolling temporal K/V cache and a world-aligned spatial cache, enabling persistent long-horizon rollouts with real-time inference at 11 FPS on a single H100 GPU. LidarVerse achieves state-of-the-art performance on most reported metrics across all three evaluation protocols, while its tokenizer supports 32-, 64-, and 128-beam LiDAR and generalizes to an unseen scanning pattern. These results demonstrate that generative and reconstruction-grounded LiDAR simulation can be realized within one model, providing the core capabilities for future closed-loop sensor simulation.

Che, Henry, Ren, Xuanchi, Fidler, Sanja · 351 views · Oct 13, 2026

LiDARWorld Models

ROS 2 3D LiDAR Localization on Prior Maps

ROS 2 3D LiDAR Localization on Prior Maps

@rsasaki0109 · 176 views · 2026-10-16

ROS2LiDAR Localization

AMB3R-SLAM: kilometer-scale real-time SLAM on a single consumer GPU, no bundle adjustment needed

AMB3R-SLAM: kilometer-scale real-time SLAM on a single consumer GPU, no bundle adjustment needed

@rsasaki0109 · 608 views · 2026-09-30

3D Climbing Analysis with iPhone LiDAR

3D Climbing Analysis with iPhone LiDAR

@jeremyparkphd · 1k views · 2026-09-24

Computer VisionLiDAR

Physical Agents SDK Turns Deployment Sites into Real2Sim Worlds

Physical Agents SDK Turns Deployment Sites into Real2Sim Worlds

@InvLambda · 629 views · 2026-09-21

Sim-to-RealDigital twins

Caltech Unitree G1 Climbs Stairs Across Floors Without a Map

Caltech Unitree G1 Climbs Stairs Across Floors Without a Map

@yokoshaan · 578 views · 2026-09-21

Legged LocomotionStair Climbing

MATRiX Closed-Loop Robot Simulation on Jetson and RK3588

MATRiX Closed-Loop Robot Simulation on Jetson and RK3588

@kikihuangyue · 748 views · 2026-09-14

Robotics SimulationProcessor-in-the-Loop

Hyundai MobED: A Production-Ready Adaptive Mobility Robot

Hyundai MobED: A Production-Ready Adaptive Mobility Robot

@spaceandtech_ · 738 views · 2026-05-03

Mobile robotsWheeled robots

Learning Agile Perceptive Traversal of Sparse 3D Structures for Humanoids

Learning Agile Perceptive Traversal of Sparse 3D Structures for Humanoids

Traversing sparse 3D structures requires humanoid robots to perceive thin, overhanging geometry while executing agile, accurate whole-body motions. We study this problem through monkey-bar traversal, where the robot must jump to the structure, traverse it through sparse bar interactions, and land safely. For this task, we present a reinforcement-learning-based perceptive control system that operates directly on observations from a head-mounted solid-state lidar. To extract task-relevant geometry from the sparse returns, the policy consumes the raw lidar scan through an attention-based encoder with recurrent memory. This policy is obtained by a phase-scheduled teacher- student pipeline that combines privileged experts for jumping up, brachiating, and jumping down. For transfer to hardware, we model lidar noise, battery-voltage sag, and actuator thermal limits, and equip the humanoid with passive hook end-effectors for robust bar interaction. On hardware, the resulting policy completes the full jump-up->brachiation->jump-down sequence in 14 of 15 trials across three bar configurations and reaches brachiation speeds up to 0.5 m/s. Beyond brachiation, the same perception backbone supports a separately trained policy that ducks beneath thin overhead obstacles with 2 cm cross-sections.

Ongan, Efe, Zhang, Chong, Sun, Boyang · 940 views · Aug 30, 2026

HumanoidBrachiation

Ouster OS1 Max Rev8 256-Channel Long-Range High-Resolution Imaging Lidar

Ouster OS1 Max Rev8 256-Channel Long-Range High-Resolution Imaging Lidar

The Ouster OS1 Max is the flagship long-range imaging lidar of the Rev8 OS series and the first to reach 256 vertical channels (also 64/128): 865 nm Class 1 eye-safe laser, 360 deg x 43.9 deg FOV, 512-4096 horizontal columns, 5-40 Hz configurable rotation. In 1024 @ 10 Hz mode it reaches 350 m on 80% Lambertian targets and 200 m on 10% targets (both >90% detection probability at 100 klx sunlight), with 500 m max representable and 0 m minimum range; range accuracy is +/-1.25 cm (Lambertian) / +/-2.5 cm (retroreflective) at 0.1 cm resolution, up to 2 returns and 10,485,760 points per second. Native RGB-D color point cloud (116 dB dynamic range) carries RGB, range, signal, reflectivity, NIR, channel, azimuth and timestamp per point; a synchronous IMU runs at 640/1280/2560 Hz. Data leaves over gigabit Ethernet UDP with PTP/gPTP/NMEA/PPS time sync (<1 ms error) and <10 ms latency. It runs from 12/24 VDC (9-58 V) at 15-25 W (19 W nominal, 28 W peak starting at -40 C), measures 87 mm diameter x 82 mm and weighs 670 g (720 g with halo cap); IP68/IP69K, -40 to +85 C, 100 g shock and 10 Grms vibration. Engineered for functional safety (ASIL-B, SIL-2, PLd) with an on-sensor 3D Zone Monitor and designed for cybersecurity to ISO 21434 / UNECE WP.29 / IEC 62443 / ISO 27001. Built for long-range industrial sensing, off-highway vehicles, autonomous cars, trucks and buses, traffic and security monitoring, and 3D mapping vehicles.

COMPONENT

Ouster REV8: Native Color LiDAR, Ruggedized and Maximized

Ouster REV8: Native Color LiDAR, Ruggedized and Maximized

Ouster · 735 views · 2026-05-05

Axioma: Open-Source ROS 2 Robot for Mapping and Autonomous Navigation

Axioma: Open-Source ROS 2 Robot for Mapping and Autonomous Navigation

@Alacritic_Super · 1.4k views · 2026-09-03

Manifold Tech Odin1 Spatial Memory Module

Manifold Tech Odin1 Spatial Memory Module

Manifold Tech's Odin1 is a spatial memory module for robots and UAVs: it packs a 940 nm, Class 1 eye-safe dToF depth engine (240x180 resolution, 0.2-30 m at 10% reflectivity, up to 70 m at 90%, up to 700,000 pts/s, +/-3 cm at 1 sigma, 120x90 deg FOV, 0.5x0.5 deg angular resolution, up to 15 FPS), a 1600x1296 global-shutter RGB camera (129 H x 104 V x 173 D deg FOV) and a high-rate IMU into a ~280 g, IP66 body measuring 100x62x43 mm. On-board compute provides native space-time synchronization and runs the in-house MindSLAM stack, so multi-sensor fusion, 3D mapping, 6DOF pose estimation and spatial-memory relocalization all happen inside the module, which then streams pose, true-color/raw point clouds, raw RGB and high-frequency IMU data directly - no external high-performance computer, no hand-built localization stack, no sensor-fusion tuning. Developers get a Linux C++ SDK plus ROS 1 / ROS 2 drivers (odin_ros_driver with point-cloud-to-depth, reprojection and image-overlay nodes and RViz configs) and can deploy custom SLAM algorithms; the MindCloud Studio / Go / Render suite and MindCloudX.AI cloud platform complete a real-to-sim loop of Pocket2 capture, MindCloud 3DGS processing, Isaac Sim training and Odin1 deployment. Official figures quote +/-5 cm + 1% SLAM positioning accuracy (with RTK correction) and 0.2% drift, 9-26 V input at 11 W rated / 24 W peak, and -10 C to 40 C operation. Odin1 ships on platforms such as Unitree H1, Astrall Dynamics and RoboForce TITAN, and sells for USD 999 in the official store (SKU ODIN1-001, B2B quoting for volume).

COMPONENT

Spatial Memory ModuleSLAM

ZVISION NZ1 Wide-Angle 3D LiDAR for All Scenarios

ZVISION NZ1 Wide-Angle 3D LiDAR for All Scenarios

ZVISION's NZ-series semi-solid-state LiDAR: 240° horizontal and 80° vertical FOV, up to 150 m range (115 m per other sources), shown in 96- and 192-line configurations; its SPAD-SoC receiver plus semi-solid-state scanning is a rare combination in short-range LiDAR, targeting drones, robot dogs, autonomous sweepers, delivery robots, forklifts and mowers.

COMPONENT

Seyond Hummingbird D1R Solid-State Ultra-Wide-Angle LiDAR

Seyond Hummingbird D1R Solid-State Ultra-Wide-Angle LiDAR

Seyond's solid-state LiDAR optimized for robotics: pure electronic scanning with zero mechanical moving parts, a 140°x100° FOV and detection of obstacles as low as 10 cm — high reliability and long life for mobile robots, AMRs and autonomous forklifts.

COMPONENT

Wanji WLR-750L Solid-State LiDAR-Video Integrated Sensor

Wanji WLR-750L Solid-State LiDAR-Video Integrated Sensor

A solid-state LiDAR-video integrated sensor that simultaneously outputs 1200-class dense point clouds and 1600x1200 high-resolution RGBD, with millisecond-level synchronization and pixel-level alignment between point cloud and image — native RGB+Depth fusion, already winning design-ins from multiple embodied robotics companies.

COMPONENT

Wanji WLR-723 Non-Repetitive Scanning 100° Full-FOV LiDAR

Wanji WLR-723 Non-Repetitive Scanning 100° Full-FOV LiDAR

The world's first non-repetitive-scanning LiDAR with a 100° full FOV: vertical FOV from -10° to +90°, up to 70 m range, 319k points/s and ±2 cm precision, accumulating point clouds evenly across the FOV for embodied robots, AMRs and other large-FOV close-range perception scenarios.

COMPONENT

Livox Mid-360L Lightweight Low-Power 360° Hybrid Solid-State LiDAR

Livox Mid-360L Lightweight Low-Power 360° Hybrid Solid-State LiDAR

The latest refresh of the Livox Mid-360 family — the most widely deployed 360° hybrid solid-state LiDAR in robotics — now lighter, lower-power and more cost-optimized, with ~100k points/s point density, targeting cost- and weight-sensitive mobile robots such as service robots, AGVs/AMRs and inspection platforms.

COMPONENT

Hesai ETX 6D Full-Color LiDAR (Picasso SPAD-SoC)

Hesai ETX 6D Full-Color LiDAR (Picasso SPAD-SoC)

A 6D full-color LiDAR built on Hesai's Picasso SPAD-SoC: every pixel carries depth and RGB color simultaneously in real-time imaging, supporting up to 4320-line full-color 4K perception with native RGBD output from a single chip — pointing LiDAR from geometry sensing toward spatial vision, with mass production planned for H2 2026.

COMPONENT

As an Amazon Associate, we earn from qualifying purchases.