Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

3D Mapping of Glacier Moulins: Challenges and lessons learned

William Dubois, Matěj Boxan, Johann Laconte, François Pomerleau

TL;DR

The paper reports a field demonstration of a custom lidar-inertial mapping platform deployed inside a glacier moulin on the Athabasca Glacier to map the surrounding environment. It highlights the core challenge of ICP-based SLAM in feature-poor, dynamic glacial interiors and documents the operational hurdles and safety considerations encountered during the week-long deployment. The authors detail the sensor suite (lidar RS-16, dual IMUs, barometer) and the need to offline-process maps due to platform jitter and low feature density. They discuss procedural preparation, testing, and environmental safeguards as critical lessons, and propose hardware and sensor fusion enhancements for more robust glacial mapping in future work.

Abstract

In this paper, we present a field report of the mapping of the Athabasca Glacier, using a custom-made lidar-inertial mapping platform. With the increasing autonomy of robotics, a wider spectrum of applications emerges. Among these, the surveying of environmental areas presents arduous and hazardous challenges for human operators. Leveraging automated platforms for data collection holds the promise of unlocking new applications and a deeper comprehension of the environment. Over the course of a week-long deployment, we collected glacier data using a tailor-made measurement platform and reflected on the inherent challenges associated with such experiments. We focus on the insights gained and the forthcoming challenges that robotics must surmount to effectively map these terrains.

3D Mapping of Glacier Moulins: Challenges and lessons learned

TL;DR

The paper reports a field demonstration of a custom lidar-inertial mapping platform deployed inside a glacier moulin on the Athabasca Glacier to map the surrounding environment. It highlights the core challenge of ICP-based SLAM in feature-poor, dynamic glacial interiors and documents the operational hurdles and safety considerations encountered during the week-long deployment. The authors detail the sensor suite (lidar RS-16, dual IMUs, barometer) and the need to offline-process maps due to platform jitter and low feature density. They discuss procedural preparation, testing, and environmental safeguards as critical lessons, and propose hardware and sensor fusion enhancements for more robust glacial mapping in future work.

Abstract

In this paper, we present a field report of the mapping of the Athabasca Glacier, using a custom-made lidar-inertial mapping platform. With the increasing autonomy of robotics, a wider spectrum of applications emerges. Among these, the surveying of environmental areas presents arduous and hazardous challenges for human operators. Leveraging automated platforms for data collection holds the promise of unlocking new applications and a deeper comprehension of the environment. Over the course of a week-long deployment, we collected glacier data using a tailor-made measurement platform and reflected on the inherent challenges associated with such experiments. We focus on the insights gained and the forthcoming challenges that robotics must surmount to effectively map these terrains.
Paper Structure (4 sections, 3 figures)

This paper contains 4 sections, 3 figures.

Table of Contents

  1. Introduction
  2. Experiments
  3. Challenges and Lessons Learned
  4. Conclusion

Figures (3)

  • Figure 1: Deployment conducted on the Athabasca Glacier. The experimental platform was lowered in a glacial moulin, mapping its surroundings.
  • Figure 2: Data gathering platform with which sensor measurements were recorded to perform 3D localization and mapping, equipped with a lidar Robosense RS-16 (1), an Xsens MTi-10 IMU (2), a Vectornav vn100 IMU (3, behind the Xsens MTi-10) and a barometric pressure sensor DPS310 (4, on the other side of the platform)
  • Figure 3: Map result (colored by elevation) from the moulin experiment. Left: top view; Right: Side view. The lack of features in the moulin makes the mapping of such environments challenging.