{"success":true,"data":{"author":"Anthony Bisulco et al. (University of Pennsylvania)","bucketAccessType":"public","bucketName":"anthonytec2/OctoSense","bucketPrefix":"","category":"Robotics","cloudProvider":"huggingface","createdAt":"2026-11-14T14:15:10.787Z","description":"OctoSense is a time-synchronised, calibrated multi-sensor dataset in which a car, a boat and a Unitree Go2-W quadruped share the same sensor rig (arXiv:2606.27317). Most of it is driving: 371 sequences, 59 hours and 2,474 km of urban, suburban and rural driving on Long Island and in Philadelphia at sunrise, day, sunset and night. Every sequence carries stereo RGB, stereo event cameras, infrared, Ouster LiDAR and IMU; driving sequences add GPS, CAN, ground truth for depth, ego-motion optical flow and semantic segmentation, and a fused GPS plus LiDAR-inertial odometry trajectory. Nine boat sequences and two quadruped sequences (with per-joint robot state) have no perception ground truth. Sequences are keyed by recording datetime under car/, boat/ and unitree/, with per-sequence data in HDF5 (data.h5), calibrations/, and metadata.parquet listing platform and attributes. Measured 2026-11-14 at about 8.5 TB. Read with h5py and pandas; the authors provide a Colab getting-started notebook. Useful for event-based vision, depth and flow estimation, SLAM and night driving. Released under the MIT licence; cite the OctoSense paper.","displayTitle":"OctoSense: Synchronized Multi-Sensor Driving Dataset with Event Cameras","id":"57f45ad6-eaef-5961-91c2-ac998f9dbc68","license":"MIT","licenseEvidenceUrl":"https://huggingface.co/datasets/anthonytec2/OctoSense","objectCountEstimate":3058,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2606.27317","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"anthonytec2/OctoSense\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"anthonytec2/OctoSense\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=anthonytec2, license:mit, size_categories:n<1K, format:parquet, downloads=9605"],"format":"mixed","generatedAt":"2026-11-14T14:15:18.052Z","layout":{"example":"boat/sess1/rosbag2_2025_11_20-08_48_14/data.h5","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"boat/sess1/rosbag2_2025_11_20-08_48_14/data.h5","truncatedAt":512}],"sampleObjects":["assets/banner_inline.png","assets/cal_axis.png","assets/gt.png","assets/gt_grid.mp4","assets/gt_grid_poster.png"],"sidecars":[{"bytes":"---\npretty_name: OctoSense\nlicense: mit\ntask_categories:\n  - depth-estimation\n  - image-segmentation\n  - robotics\ntags:\n  - robotics\n  - multimodal\n  - autonomous-driving\n  - quadruped\n  - marine\n  - lidar\n  - event-camera\n  - stereo\n  - optical-flow\n  - ego-motion\n  - sensor-fusion\n  - infrared\n  - gps\n  - geospatial\n  - 3d\n  - timeseries\nsize_categories:\n  - n>1T\nconfigs:\n  - config_name: default\n    data_files: metadata.parquet\n---\n<p align=\"center\">\n  <img src=\"assets/banner_inline.png\" alt=\"OctoSense\" width=\"760\"/>\n</p>\n\n<p align=\"center\">\n  <a href=\"https://arxiv.org/abs/2606.27317\"><img src=\"https://img.shields.io/badge/arXiv-2606.27317-b31b1b?logo=arxiv\" alt=\"arXiv\"/></a>\n  <a href=\"https://github.com/anthonytec2/OctoSense\"><img src=\"https://img.shields.io/badge/Code-GitHub-181717?logo=github\" alt=\"Code on GitHub\"/></a>\n  <a href=\"https://colab.research.google.com/drive/11hBm5dH--7HueAFCKtaAFnDAJul-52fb?usp=sharing\">\n    <img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open the quickstart in Colab\"/>\n  </a>\n</p>\n\n> **New to OctoSense?** A [getting-started Colab notebook](https://colab.research.google.com/drive/11hBm5dH--7HueAFCKtaAFnDAJul-52fb?usp=sharing) walks through loading a sequence and using each modality. Click **Open in Colab** above to run it in your browser, no setup required.\n\n<p align=\"center\">\n  <video autoplay loop muted playsinline width=\"100%\" poster=\"https://huggingface.co/datasets/anthonytec2/OctoSense/resolve/main/assets/hero.gif\">\n    <source src=\"https://huggingface.co/datasets/anthonytec2/OctoSense/resolve/main/assets/hero.mp4\" type=\"video/mp4\"/>\n    <img src=\"assets/hero.gif\" alt=\"OctoSense, synchronized multi-sensor driving across day, night, and degraded conditions\" width=\"100%\"/>\n  </video>\n</p>\n\n**OctoSense** is a time-synchronized, calibrated, multi-sensor dataset spanning multiple platforms, all sharing the same sensor rig. The bulk of the data is large-scale **driving dataset**: **371 sequences · 59 hrs · 2,474 km · 8.43 TB** of urban, suburban, and rural driving (highway / residential / city) on **Long Island** and in **Philadelphia**, across sunrise, daytime, sunset, and nighttime, complemented by additional sequences captured on a **boat** and a **Unitree quadruped**. Every sequence carries synchronized stereo RGB, stereo event cameras, infrared, Ouster LiDAR, and IMU; the driving sequences additionally provide GPS, CAN, ground truth (depth, ego-motion optical flow, semantic segmentation), and a fused GPS + LiDAR-inertial odometry trajectory. Each sequence's recording **`platform`** (`car` / `boat` / `unitree`) is a column in `metadata.parquet`: see [Additional platforms](#additional-platforms--boat--unitree) for the boat/unitree specifics.\n\n\n> Canonical key for every sequence is its recording datetime, `rosbag2_YYYY_MM_DD-HH_MM_SS`.\n> Per-sequence attributes live in **`metadata.parquet`** (the viewer's table: a sensor-montage thumbnail, a GPS-track map, and all metadata columns per sequence).\n\n## Additional platforms — Boat & Unitree\n\n<p align=\"center\">\n  <video autoplay loop muted playsinline width=\"100%\" poster=\"https://huggingface.co/datasets/anthonytec2/OctoSense/resolve/main/assets/platforms_poster.png\">\n    <source src=\"https://huggingface.co/datasets/anthonytec2/OctoSense/resolve/main/assets/platforms.mp4\" type=\"video/mp4\"/>\n    <img src=\"assets/platforms_poster.png\" alt=\"Boat and Unitree-quadruple sensor data (synchronized RGB, event, infrared, LiDAR)\" width=\"100%\"/>\n  </video>\n</p>\n\nAlongside the 371 **car/driving** sequences, OctoSense includes sequences from two other platforms recorded with the same sensor rig, distinguished by the **`platform`** column in `metadata.parquet` (`car` / `boat` / `unitree`). No perception ground truth is provided for these platforms. \n\n- **`boat/sess1/`: 9 sequences**: No LiDAR odometry or fused trajectory, and the LiDAR is raw rather than motion-compensated: `data.h5` stores the raw Ouster `range`/`signal`/`ir`/`reflectivity` images instead of the deskewed `ouster/range_pcl` point cloud. Note the boat data u","encoding":"utf8","key":"README.md","name":"README.md","size":25739,"truncated":true},{"bytes":"[{\"window_id\": 0, \"video_id\": \"sess10/rosbag2_2026_01_21-07_35_29\", \"frame_idx\": 0, \"timestamp\": 0.0009782214220179137, \"caption\": \"An urban street with two lanes and a white dashed centerline is lined with multi-story brick and concrete buildings. A white convertible and several other parked cars line the right curb. In the distance, several vehicles travel in both directions, including a white car moving away in the right lane. A green street sign reads \\\"ST JAMES\\\". The road surface is dry under clear dawn light. The ego vehicle is decelerating.\", \"speed_mps\": 8.74942398071289, \"turn_deg\": 0.5145962238311768, \"dist_m\": 37.73241424560547, \"is_night\": 1}, {\"window_id\": 1, \"video_id\": \"sess10/rosbag2_2026_01_21-07_35_29\", \"frame_idx\": 500, \"timestamp\": 5.0010241393608075, \"caption\": \"An urban street with two lanes and white dashed lane markings is flanked by multi-story brick and modern buildings. A silver SUV in the left lane and a silver sedan in the right lane are at mid-distance, both moving forward. A white hatchback is parked on the right shoulder near the foreground. Further ahead, several vehicles are queued in both lanes. A yellow diamond-shaped sign is visible on the left, and a sign on the right reads \\\"CHANCELLOR\\\". The road surface is dry under clear dawn lighting. The ego vehicle is decelerating.\", \"speed_mps\": 6.219781398773193, \"turn_deg\": 0.5590729117393494, \"dist_m\": 31.721416473388672, \"is_night\": 1}, {\"window_id\": 2, \"video_id\": \"sess10/rosbag2_2026_01_21-07_35_29\", \"frame_idx\": 1000, \"timestamp\": 10.001017323660715, \"caption\": \"An urban street with two lanes and white lane markings is flanked by multi-story buildings. A silver SUV is in the left lane at mid-distance, while a silver sedan is centered in the right lane at mid-distance. Further ahead in the left lane is a white van, and a silver SUV is parked or idling on the right side of the road. A white car is parked on the right shoulder in the foreground. The road surface is dry under clear dawn light. The ego vehicle decelerates from 3 to 0 m/s.\", \"speed_mps\": 0.7799388766288757, \"turn_deg\": 0.4834608733654022, \"dist_m\": 3.977541208267212, \"is_night\": 1}, {\"window_id\": 3, \"video_id\": \"sess10/rosbag2_2026_01_21-07_35_29\", \"frame_idx\": 1500, \"timestamp\": 15.001020870312004, \"caption\": \"An urban street with a single white lane marking is crowded with traffic during dawn. A silver SUV is positioned near in the left lane, and a silver sedan is directly ahead at mid-distance. Further ahead in the center lane is a white van, while several silver and white vehicles are parked or idling along the right curb. The road surface is dry. Tall buildings line both sides of the street, with the morning light creating glare on the distant skyline. The ego vehicle remains stationary.\", \"speed_mps\": 0.06689108163118362, \"turn_deg\": 0.20963124930858612, \"dist_m\": 0.34115758538246155, \"is_night\": 1}, {\"window_id\": 4, \"video_id\": \"sess10/rosbag2_2026_01_21-07_35_29\", \"frame_idx\": 2000, \"timestamp\": 20.001013739569455, \"caption\": \"An urban street with a single white lane marking is lined with multi-story buildings under dawn light. A silver SUV is stopped in the left lane at near distance. A silver sedan is stopped in the center lane at mid-distance with its taillights illuminated. Further ahead in the center lane, a white van is stopped. To the right, a silver SUV and a white sedan are parked along the curb. The road surface is dry. The ego vehicle remains stationary.\", \"speed_mps\": 0.0594334751367569, \"turn_deg\": 0.23060190677642822, \"dist_m\": 0.3031032085418701, \"is_night\": 1}, {\"window_id\": 5, \"video_id\": \"sess10/rosbag2_2026_01_21-07_35_29\", \"frame_idx\": 2500, \"timestamp\": 25.00101155556057, \"caption\": \"An urban street with a single white lane marking is crowded with traffic during dawn. A silver SUV is positioned near in the left lane, and a silver sedan is directly ahead at mid-distance with its taillights illuminated. Further ahead in the center lane is a white van, and several other vehicles are queued in the distance. To the right, a silver SUV and a white sed","encoding":"utf8","key":"semantic_search/metadata.json","name":"metadata.json","size":30372841,"truncated":true}],"sourceObjectKey":"boat/sess1/rosbag2_2025_11_20-08_48_14/data.h5","starterCell":"from huggingface_hub import hf_hub_download\npath = hf_hub_download(repo_id=\"anthonytec2/OctoSense\", filename=\"boat/sess1/rosbag2_2025_11_20-08_48_14/data.h5\", repo_type=\"dataset\")\nprint(path)","starterSamplePath":"assets/banner_inline.png","totalObjects":3058},"summary":"OctoSense is a time-synchronised, calibrated multi-sensor dataset in which a car, a boat and a Unitree Go2-W quadruped share the same sensor rig (arXiv:2606.27317).","title":"OctoSense: Synchronized Multi-Sensor Driving Dataset with Event Cameras","updatedAt":"2026-11-14T17:43:01.484Z","visibility":"public","websiteUrl":"https://github.com/anthonytec2/OctoSense"}}