Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

GO: The Great Outdoors Multimodal Dataset

Peng Jiang, Kasi Viswanath, Akhil Nagariya, George Chustz, Maggie Wigness, Philip Osteen, Timothy Overbye, Christian Ellis, Long Quang, Srikanth Saripalli

TL;DR

The GO dataset addresses the need for robust perception and navigation in unstructured, off-road environments by providing a comprehensive multimodal resource. It combines a wide sensor suite (LiDAR, radar, thermal, RGB, stereo, IMU, GNSS, RTK GPS) with precise synchronization and centimeter-precision ground-truth trajectories, enabling advanced tasks such as semantic segmentation, object detection, and SLAM. A 22-class ontology, semi-automated yet human-refined annotations, and five teleoperated routes totaling 10.26 km across diverse terrains constitute the core contributions. The work outlines research directions in robust localization, multi-modal perception, and navigation, and highlights open questions on generalization, sensor fusion efficacy, and standardized evaluation—establishing GO as a valuable benchmark for autonomous field robotics.

Abstract

The Great Outdoors (GO) dataset is a multi-modal annotated data resource aimed at advancing ground robotics research in unstructured environments. This dataset provides the most comprehensive set of data modalities and annotations compared to existing off-road datasets. In total, the GO dataset includes six unique sensor types with high-quality semantic annotations and GPS traces to support tasks such as semantic segmentation, object detection, and SLAM. The diverse environmental conditions represented in the dataset present significant real-world challenges that provide opportunities to develop more robust solutions to support the continued advancement of field robotics, autonomous exploration, and perception systems in natural environments. The dataset can be downloaded at: https://www.unmannedlab.org/the-great-outdoors-dataset/

GO: The Great Outdoors Multimodal Dataset

TL;DR

The GO dataset addresses the need for robust perception and navigation in unstructured, off-road environments by providing a comprehensive multimodal resource. It combines a wide sensor suite (LiDAR, radar, thermal, RGB, stereo, IMU, GNSS, RTK GPS) with precise synchronization and centimeter-precision ground-truth trajectories, enabling advanced tasks such as semantic segmentation, object detection, and SLAM. A 22-class ontology, semi-automated yet human-refined annotations, and five teleoperated routes totaling 10.26 km across diverse terrains constitute the core contributions. The work outlines research directions in robust localization, multi-modal perception, and navigation, and highlights open questions on generalization, sensor fusion efficacy, and standardized evaluation—establishing GO as a valuable benchmark for autonomous field robotics.

Abstract

The Great Outdoors (GO) dataset is a multi-modal annotated data resource aimed at advancing ground robotics research in unstructured environments. This dataset provides the most comprehensive set of data modalities and annotations compared to existing off-road datasets. In total, the GO dataset includes six unique sensor types with high-quality semantic annotations and GPS traces to support tasks such as semantic segmentation, object detection, and SLAM. The diverse environmental conditions represented in the dataset present significant real-world challenges that provide opportunities to develop more robust solutions to support the continued advancement of field robotics, autonomous exploration, and perception systems in natural environments. The dataset can be downloaded at: https://www.unmannedlab.org/the-great-outdoors-dataset/

Paper Structure

This paper contains 12 sections, 5 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Dataset Descriptions
  3. Sensor Setup
  4. Sensor Calibration and Synchronization
  5. Route Information
  6. Semantic Annotations:
  7. Ontology
  8. Annotation process
  9. Class Distribution
  10. Research Directions and Open Questions
  11. Conclusion
  12. Biography Section

Figures (5)

  • Figure 1: The GO dataset comprises five routes, covering a cumulative distance of 10.26 km and a total duration of 98.60 minutes.
  • Figure 2: Example of the raw perception sensor data from the GO Dataset. The figure shows (a) the left stereo camera view, (b) the right stereo camera view, (c) the rear camera view, (d) thermal camera imagery, and (e) a combined representation of LiDAR (white) and threshold-filtered radar data (color)
  • Figure 3: Qualitative visualization of the LiDAR-Camera calibration.
  • Figure 4: Example images and semantic segmentation labels from the GO Dataset. The figure shows (a) a forest area and (b) a trail, along with their respective semantic segmentation results.
  • Figure 5: Image Label distribution. The tree, grass, dirt and sky constitute the major classes.