Geppetto tracks 1035 robot training datasets including Open X-Embodiment, DROID, and AgiBot World, each with structured metadata on trajectory count, hours of demonstration, robot platforms, license, commercial use, access model, and file format. Every dataset connects to the robots it trains and the professions it affects.
| Dataset | Year | Category | Trajectories | Hours | Platforms | License | Commercial | Access | Format |
|---|---|---|---|---|---|---|---|---|---|
| ORBIT | 2021 | human-video | — | — | mobile phones held by blind or low-vision users | not stated | No | open | — |
| TREK-150 | 2021 | human-video | — | — | EPIC-KITCHENS videos | not stated | No | open | — |
| H2O-3D | 2021 | human-video | — | — | fixed multi-view RGB-D cameras (not egocentric) | not stated | No | open | — |
| ARQ-CRISP slip detection dataset 2021 | 2021 | research-lab | — | — | robot hand tactile data during pick and place | not stated | No | open | — |
| Robot handle grasping failure data | 2021 | research-lab | — | — | Toyota HSR drawer handle grasping | CC-BY-4.0 | Yes | open | — |
| Grasps Under Varied Object Orientation Dataset | 2021 | research-lab | — | — | human hand grasps | not stated | No | open | — |
| OCID-Grasp | 2021 | research-lab | — | — | OCID RGB-D scenes with grasp rectangles | see repo LICENSE | No | open | — |
| SuctionNet-1Billion | 2021 | research-lab | — | — | GraspNet-1Billion scenes with suction labels | CC BY-NC-SA 4.0 | No | open | — |
| Contact-GraspNet training scenes | 2021 | research-lab | — | — | simulated tabletop scenes, parallel-jaw | see repo License.pdf | No | open | — |
| DefGraspSim | 2021 | research-lab | — | — | simulated grasps on deformable objects | see repo LICENSE | No | open | — |
| TrajAir | 2021 | real-robot | — | 2,442 | drone | CC-BY-4.0 | Yes | open | — |
| DexYCB | 2021 | human-video | — | 0.83 | human | CC-BY-NC-4.0 | No | open | — |
| In-flight Positional and Energy-Use Dataset of Package-Delivery Quadcopter UAVs | 2021 | real-robot | — | 10.75 | drone | CC-BY-4.0 | Yes | open | — |
| Anti-UAV | 2021 | research-lab | — | — | RGB and thermal video of drones | not stated | No | open | — |
| CrossLoc | 2021 | research-lab | — | — | real and rendered aerial imagery | not stated | No | open | — |
| Fixed-wing Aircraft Collision Course Dataset | 2021 | research-lab | — | — | camera, fixed-wing aircraft | not stated | No | open | — |
| Lighthouse Positioning System Dataset | 2021 | research-lab | — | — | Lighthouse tracking for UAVs | not stated | No | open | — |
| NeuroBEM dataset | 2021 | research-lab | — | — | quadrotor in a large motion capture system | not stated | No | open | — |
| IPS300+ | 2021 | research-lab | — | — | roadside lidar and cameras at intersection | CC BY-NC-SA 4.0 | No | open | — |
| REGRAD | 2021 | research-lab | — | — | simulated cluttered scenes with relationship and grasp labels | not stated | No | open | — |
| UAV-Human | 2021 | research-lab | — | — | UAV multi-modal video of people | not stated | No | open | — |
| ExPI (Extreme Pose Interaction) | 2021 | research-lab | — | — | two dancers, 3D poses and shapes | not stated | No | open | — |
| SAMP | 2021 | research-lab | — | — | human body interacting with furniture | not stated | No | open | — |
| Airbert BnB | 2021 | research-lab | — | — | online rental listing images | not stated | No | open | — |
| OCID-Ref | 2021 | research-lab | — | — | OCID RGB and point cloud scenes | MIT | Yes | open | — |
| ScanQA | 2021 | research-lab | — | — | ScanNet RGB-D indoor scans | CC BY-NC-SA 3.0 | No | open | — |
| SUTD-TrafficQA | 2021 | research-lab | — | — | in-the-wild traffic videos | not stated | No | open | — |
| K-EQA | 2021 | research-lab | — | — | simulated scenes with external knowledge | not stated | No | open | — |
| Robo-VLN | 2021 | research-lab | — | — | continuous 3D reconstructed environments | not stated | No | open | — |
| SOON | 2021 | research-lab | — | — | Matterport3D simulator | not stated | No | open | — |
| MIT Mini Cheetah contact data sets | 2021 | research-lab | — | — | MIT Mini Cheetah proprioception | not stated | No | open | — |
| AcinoSet | 2021 | research-lab | — | — | cheetahs in the wild, multi-view | not stated | No | open | — |
| ColoRadar | 2021 | research-lab | — | — | millimeter wave radar rig | not stated | No | open | — |
| Hilti SLAM Challenge Dataset 2021 | 2021 | research-lab | — | — | handheld multi-sensor rig | not stated | No | open | — |
| TUM-VIE | 2021 | research-lab | — | — | stereo event cameras, stereo grayscale, IMU | not stated | No | open | — |
| CMU-GPR | 2021 | research-lab | — | — | ground penetrating radar | not stated | No | open | — |
| M2DGR | 2021 | research-lab | — | — | ground robot with fisheye, infrared, event cameras, LiDAR, IMU | not stated | No | open | — |
| NAVER LABS indoor localization datasets | 2021 | research-lab | — | — | mapping platform with 10 cameras and LiDAR | not stated | No | open | — |
| EgoCom | 2020 | research-lab | — | 38.5 | — | Modified MIT | Yes | open | — |
| SensatUrban | 2020 | research-lab | — | — | urban-scale photogrammetric point clouds | not stated | No | open | — |
| University-1652 | 2020 | research-lab | — | — | synthetic drone, satellite and ground views | not stated | No | open | — |
| pNEUMA | 2020 | research-lab | — | — | swarm of 10 drones over Athens | not stated | No | open | — |
| 4Seasons | 2020 | research-lab | — | — | car with stereo camera, IMU, RTK-GNSS | CC BY-NC-SA 4.0 | No | open | — |
| A2D2 | 2020 | research-lab | — | — | Audi Q7 with 6 cameras, 5 lidars | CC BY-ND 4.0 | No | open | — |
| Car Crash Dataset (CCD) | 2020 | research-lab | — | — | dashcam accident videos | not stated | No | open | — |
| CARRADA | 2020 | research-lab | — | — | camera and automotive radar | CC BY-NC-SA 4.0 | No | open | — |
| Cirrus | 2020 | research-lab | — | — | car with two 250 m bi-pattern lidars and camera | not stated | No | open | — |
| comma10k | 2020 | research-lab | — | — | comma device road images | MIT | Yes | open | — |
| CurveLanes | 2020 | research-lab | — | — | lane images | not stated | No | open | — |
| DDAD | 2020 | research-lab | — | — | TRI vehicles, 6 cameras, Luminar lidar | CC BY-NC-SA 4.0 | No | open | — |
| Ford Multi-AV Seasonal Dataset | 2020 | research-lab | — | — | Ford AV fleet, lidars, cameras, GPS/IMU | CC BY-NC-SA 4.0 | No | open | — |
| FSOCO | 2020 | research-lab | — | — | Formula Student driverless car cameras | not stated | No | open | — |
| Honda TITAN | 2020 | research-lab | — | — | egocentric vehicle camera, Tokyo | research only | No | open | — |
| LIBRE | 2020 | research-lab | — | — | multiple 3D LiDARs on vehicle | not stated | No | open | — |
| nuImages | 2020 | research-lab | — | — | nuScenes vehicle cameras | CC BY-NC-SA 4.0 | No | gated | — |
| Oxford Radar RobotCar | 2020 | research-lab | — | — | RobotCar with Navtech radar, lidar, cameras | CC BY-NC-SA 4.0 | No | gated | — |
| PandaSet | 2020 | research-lab | — | — | car with Hesai Pandar64 and PandarGT lidars, 6 cameras | CC BY 4.0 | Yes | open | — |
| PePScenes | 2020 | research-lab | — | — | pedestrian action labels in 3D | not stated | No | open | — |
| SemanticPOSS | 2020 | research-lab | — | — | LiDAR point clouds | not stated | No | open | — |
| Toronto-3D | 2020 | research-lab | — | — | mobile LiDAR urban point cloud | not stated | No | open | — |
| Ego2Hands / Ego2HandsPose | 2020 | human-video | — | — | egocentric RGB camera | not stated | No | open | — |
| EgoK360 | 2020 | human-video | — | — | wearable 360 camera | not stated | No | open | — |
| Ego3DHands | 2020 | human-video | — | — | Synthetic two-hand images | not stated | No | open | — |
| ContactPose | 2020 | research-lab | — | — | human hand grasps, RGB-D and contact maps | not stated | No | open | — |
| EGAD | 2020 | research-lab | — | — | 3D-printable object meshes | not stated | No | open | — |
| GRAB | 2020 | research-lab | — | — | human whole body with hands and face, SMPL-X, plus 3D objects | custom, sign-in required | No | gated | — |
| Event-Stream grasping dataset | 2020 | research-lab | — | — | neuromorphic event camera | not stated | No | open | — |
| GraspNet-1Billion | 2020 | research-lab | — | — | Kinect Azure and RealSense D435 RGB-D, parallel-jaw grasp labels | CC BY-NC-SA 4.0 | No | open | — |
| Real Embodied Dataset (RED) | 2020 | research-lab | — | — | hand-mounted RGB-D, upper hemisphere viewpoints | not stated | No | open | — |
| TaskGrasp | 2020 | research-lab | — | — | RGB-D point clouds of real objects with task labels | not stated | No | open | — |
| ACRONYM | 2020 | research-lab | — | — | simulated parallel-jaw gripper (NVIDIA FleX) | MIT (repo) | No | open | — |
| Robotic Interestingness Dataset (SubTF) | 2020 | human-video | — | 7.8 | — | unknown | No | open | — |
| TartanAir | 2020 | synthetic | — | 27.8 | drone | CC-BY-4.0 | Yes | open | — |
| AU-AIR | 2020 | research-lab | — | — | UAV with camera and flight telemetry | not stated | No | open | — |
| ERA | 2020 | research-lab | — | — | aerial event videos | not stated | No | open | — |
| CADC (Canadian Adverse Driving Conditions) | 2020 | research-lab | — | — | Lincoln MKZ with 8 cameras, lidar, GNSS | CC BY-NC 4.0 | No | open | — |
| VGN synthetic grasp data | 2020 | research-lab | — | — | simulated parallel-jaw grasps in clutter | BSD-3-Clause (repo) | No | open | — |
| HumanAct12 | 2020 | research-lab | — | — | human body, 3D skeleton | not stated | No | open | — |
| AIR-Act2Act | 2020 | research-lab | — | — | human-human interaction for social robots | not stated | No | open | — |
| ETRI-Activity3D | 2020 | research-lab | — | — | RGB-D from robot viewpoint | not stated | No | open | — |
| InterHand2.6M | 2020 | research-lab | — | — | multi-view interacting hands | not stated | No | open | — |
| MoGaze | 2020 | research-lab | — | — | full-body mocap with eye gaze | not stated | No | open | — |
| SFU-Store-Nav | 2020 | research-lab | — | — | camera and mocap of shoppers | not stated | No | open | — |
| GoLD (Grounded Language Dataset) | 2020 | research-lab | — | — | multimodal object descriptions | not stated | No | open | — |
| RxR (Room-Across-Room) | 2020 | research-lab | — | — | Matterport3D simulator | CC-BY-4.0 (annotations) | Yes | open | — |
| VLN-CE (R2R-CE and RxR-CE) | 2020 | research-lab | — | — | Habitat continuous environments | not stated | No | open | — |
| BDD-OIA | 2020 | research-lab | — | — | BDD100K images | not stated | No | open | — |
| RobotSlang | 2020 | research-lab | — | — | physical robot driven by a human Driver | not stated | No | open | — |
| Motion Imitation retargeted dog motions (Peng et al.) | 2020 | research-lab | — | — | Laikago and A1 quadrupeds | not stated | No | open | — |
| RGBD-Dog | 2020 | research-lab | — | — | dogs of several breeds, Kinect v2 plus mocap skeleton | not stated | No | open | — |
| Newer College Dataset | 2020 | research-lab | — | — | handheld LiDAR, inertial and vision device | not stated | No | open | — |
| TrajNet++ | 2020 | research-lab | — | — | pedestrian trajectories in crowds | not stated | No | open | — |
| Meta-World — Multi-Task and Meta-Reinforcement Learning | 2019 | synthetic | — | — | Sawyer (simulated) | MIT | Yes | open | json, pkl |
| CITR and DUT | 2019 | research-lab | — | — | top-view pedestrian and vehicle trajectories | not stated | No | open | — |
| Mapillary Traffic Sign Dataset (MTSD) | 2019 | research-lab | — | — | crowdsourced street-level cameras | CC BY-NC-SA | No | gated | — |
| A*3D | 2019 | research-lab | — | — | car with cameras and lidar, Singapore | non-commercial academic | No | open | — |
| A3D (AnAn Accident Detection) | 2019 | research-lab | — | — | dashcam accident videos | not stated | No | open | — |
| Argoverse 1 | 2019 | research-lab | — | — | Argo AI vehicle, ring cameras, lidar, HD maps | CC BY-NC-SA 4.0 | No | open | — |
| BLVD | 2019 | research-lab | — | — | vehicle, 5D semantics | not stated | No | open | — |
| Brno Urban Dataset | 2019 | research-lab | — | — | car with cameras, lidars, thermal camera, GNSS | not stated | No | open | — |