Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Collecting Larg-Scale Robotic Datasets on a High-Speed Mobile Platform

Yuxin Lin, Jiaxuan Ma, Sizhe Gu, Jipeng Kong, Bowen Xu, Xiting Zhao, Dengji Zhao, Wenhan Cao, Sören Schwertfeger

TL;DR

The paper tackles the challenge of collecting large-scale outdoor robotic datasets by addressing the speed limitations of indoor-oriented mapping robots. It proposes a high-speed data-collection platform that attaches the ShanghaiTech Mapping Robot to a hydraulic lifting flatbed car, complemented by a secure mounting system, dual-wheel odometry encoders, and an external downlooking sensor plate. The authors demonstrate the approach with a 10.21 km, 47-minute dataset collected over campus and underground parking, along with a 3D map built using LIO-SAM from a front LiDAR, totaling 73 million points in a $0.1\,$m$ resolution$. They also provide CAD files, code, and ROS data pipelines to facilitate reuse. This platform enables scalable outdoor SLAM benchmarking and large-scale data accumulation, opening avenues for richer multimodal datasets and broader benchmarking across robotics datasets.

Abstract

Mobile robotics datasets are essential for research on robotics, for example for research on Simultaneous Localization and Mapping (SLAM). Therefore the ShanghaiTech Mapping Robot was constructed, that features a multitude high-performance sensors and a 16-node cluster to collect all this data. That robot is based on a Clearpath Husky mobile base with a maximum speed of 1 meter per second. This is fine for indoor datasets, but to collect large-scale outdoor datasets a faster platform is needed. This system paper introduces our high-speed mobile platform for data collection. The mapping robot is secured on the rear-steered flatbed car with maximum field of view. Additionally two encoders collect odometry data from two of the car wheels and an external sensor plate houses a downlooking RGB and event camera. With this setup a dataset of more than 10km in the underground parking garage and the outside of our campus was collected and is published with this paper.

Collecting Larg-Scale Robotic Datasets on a High-Speed Mobile Platform

TL;DR

The paper tackles the challenge of collecting large-scale outdoor robotic datasets by addressing the speed limitations of indoor-oriented mapping robots. It proposes a high-speed data-collection platform that attaches the ShanghaiTech Mapping Robot to a hydraulic lifting flatbed car, complemented by a secure mounting system, dual-wheel odometry encoders, and an external downlooking sensor plate. The authors demonstrate the approach with a 10.21 km, 47-minute dataset collected over campus and underground parking, along with a 3D map built using LIO-SAM from a front LiDAR, totaling 73 million points in a m. They also provide CAD files, code, and ROS data pipelines to facilitate reuse. This platform enables scalable outdoor SLAM benchmarking and large-scale data accumulation, opening avenues for richer multimodal datasets and broader benchmarking across robotics datasets.

Abstract

Mobile robotics datasets are essential for research on robotics, for example for research on Simultaneous Localization and Mapping (SLAM). Therefore the ShanghaiTech Mapping Robot was constructed, that features a multitude high-performance sensors and a 16-node cluster to collect all this data. That robot is based on a Clearpath Husky mobile base with a maximum speed of 1 meter per second. This is fine for indoor datasets, but to collect large-scale outdoor datasets a faster platform is needed. This system paper introduces our high-speed mobile platform for data collection. The mapping robot is secured on the rear-steered flatbed car with maximum field of view. Additionally two encoders collect odometry data from two of the car wheels and an external sensor plate houses a downlooking RGB and event camera. With this setup a dataset of more than 10km in the underground parking garage and the outside of our campus was collected and is published with this paper.
Paper Structure (11 sections, 2 equations, 10 figures, 1 table)

This paper contains 11 sections, 2 equations, 10 figures, 1 table.

Table of Contents

  1. Introduction
  2. State of the Art
  3. System Description
  4. Mounting System
  5. Wheel Odometry
  6. Odometry Model
  7. Odometry Experiment
  8. Downlooking Sensor Plate
  9. Dataset
  10. Dataset SLAM Experiment
  11. Conclusions

Figures (10)

  • Figure 1: Flatbed car with ShanghaiTech Mapping Robot, encoders and extra downlooking cameras.
  • Figure 2: The Hydraulic lifting flatbed car.
  • Figure 3: Explode vision. a) UGV. b) Long bolts. c) wood plate. d) wooden strips. e) wooden beam.
  • Figure 4: Car with frame prototype.
  • Figure 5: The scheme for installing a bracket on the wheel. a) Adding a bracket at the center of the wheel. b) Locking the rotation of the wheel and the encoder shaft through the bracket and a flange coupling.
  • ...and 5 more figures