Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ExelMap: Explainable Element-based HD-Map Change Detection and Update

Lena Wild, Ludvig Ericson, Rafael Valencia, Patric Jensfelt

TL;DR

This paper presents ExelMap, an explainable element-based map updating strategy that specifically identifies changed map elements and discusses how currently used metrics fail to capture change detection performance, while allowing for unfair comparison between prior-less and prior-informed map generation methods.

Abstract

Acquisition and maintenance are central problems in deploying high-definition (HD) maps for autonomous driving, with two lines of research prevalent in current literature: Online HD map generation and HD map change detection. However, the generated map's quality is currently insufficient for safe deployment, and many change detection approaches fail to precisely localize and extract the changed map elements, hence lacking explainability and hindering a potential fleet-based cooperative HD map update. In this paper, we propose the novel task of explainable element-based HD map change detection and update. In extending recent approaches that use online mapping techniques informed with an outdated map prior for HD map updating, we present ExelMap, an explainable element-based map updating strategy that specifically identifies changed map elements. In this context, we discuss how currently used metrics fail to capture change detection performance, while allowing for unfair comparison between prior-less and prior-informed map generation methods. Finally, we present an experimental study on real-world changes related to pedestrian crossings of the Argoverse 2 Map Change Dataset. To the best of our knowledge, this is the first comprehensive problem investigation of real-world end-to-end element-based HD map change detection and update, and ExelMap the first proposed solution.

ExelMap: Explainable Element-based HD-Map Change Detection and Update

TL;DR

This paper presents ExelMap, an explainable element-based map updating strategy that specifically identifies changed map elements and discusses how currently used metrics fail to capture change detection performance, while allowing for unfair comparison between prior-less and prior-informed map generation methods.

Abstract

Acquisition and maintenance are central problems in deploying high-definition (HD) maps for autonomous driving, with two lines of research prevalent in current literature: Online HD map generation and HD map change detection. However, the generated map's quality is currently insufficient for safe deployment, and many change detection approaches fail to precisely localize and extract the changed map elements, hence lacking explainability and hindering a potential fleet-based cooperative HD map update. In this paper, we propose the novel task of explainable element-based HD map change detection and update. In extending recent approaches that use online mapping techniques informed with an outdated map prior for HD map updating, we present ExelMap, an explainable element-based map updating strategy that specifically identifies changed map elements. In this context, we discuss how currently used metrics fail to capture change detection performance, while allowing for unfair comparison between prior-less and prior-informed map generation methods. Finally, we present an experimental study on real-world changes related to pedestrian crossings of the Argoverse 2 Map Change Dataset. To the best of our knowledge, this is the first comprehensive problem investigation of real-world end-to-end element-based HD map change detection and update, and ExelMap the first proposed solution.
Paper Structure (14 sections, 9 equations, 3 figures, 1 table)

This paper contains 14 sections, 9 equations, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related work
  3. HD-Map Change Detection
  4. HD-Map Generation
  5. Prior-aided HD-Map Update
  6. Methodology
  7. Stale HD-Map Encoder
  8. Encoded Map Prior Integration
  9. Change Detection Heads
  10. Datasets and Synthetic Change Generation
  11. Experiments
  12. Explainable Change Detection Metrics
  13. Discussion
  14. Conclusions and Future Work

Figures (3)

  • Figure 1: Architecture of ExelMap. The HD map encoder is based on smerf with our own extensions. In addition to the standard map generation heads, the insertion and deletion heads detect changes element-wise. The output of the network is twofold: an updated local HD map, and a second map containing the element-wise change score.
  • Figure 2: Evaluation Strategies for explainable element-based Change Detection. Yellow highlights represent change-type agnostic scores, red highlights indicate lane segment deletions, green highlights indicate pedestrian crossing insertions. We differentiate between single frame (SF) and multi-frame evaluation. A mathematical description of (a)--(i) is provided in \ref{['sec:mathe']}.
  • Figure 3: Examples of element-based change detection and update on selected scenes of the Argoverse 2 Map Change Detection Dataset Argoverse2. The figures show (from left to right): HD map prior, camera images, change map (GT and ExelMap) and updated map (GT and ExelMap). In the change map, insertions are color-coded green, deletions dashed-red and unchanged elements grey.