Skip to main navigation Skip to search Skip to main content

Creating Geopositioned 3D Areas of Interest from Fleet Gaze Data

  • Volkswagen AG
  • University of Tübingen

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In order to observe driver’s attention levels, different approaches are followed. They include simple methods counting driver input changes [6], machine learning based approaches based on driver input [17], and methods considering additional inputs such as environmental data and eye tracking data [3–5, 7, 12, 16]. Recent studies have proposed geopositioned 3D AOIs as a tool for driver intention observation. Geopositioned 3D AOIs are three dimensional Areas (boxes), with fix geopositiones (e.g. GPS) which have to be observed for a safe completion of driving maneuvers. Examples are pedestrian waiting areas, crosswalks, and traffic light. Creating these AOIs by hand is a tedious task with ample room for potential errors, as the created AOIs might differ from the real AOIs drivers look at. We therefore propose a pipeline to generate real 3D AOIs from gaze clouds. To generate relevant gaze clouds we use the points of closest encounter in fleet gaze data collected in a driving simulator setup. The results show that the generation of 3D AOIs from fleet data is possible and the created AOIs are mostly consistent with the expected AOIs.

Original languageEnglish
Title of host publicationHCI in Mobility, Transport, and Automotive Systems - 4th International Conference, MobiTAS 2022, Held as Part of the 24th HCI International Conference, HCII 2022, Proceedings
EditorsHeidi Krömker
PublisherSpringer Science and Business Media Deutschland GmbH
Pages21-34
Number of pages14
ISBN (Print)9783031049866
DOIs
StatePublished - 2022
Externally publishedYes
Event4th International Conference on HCI in Mobility, Transport, and Automotive Systems, MobiTAS 2022 Held as Part of the 24th HCI International Conference, HCII 2022 - Virtual, Online
Duration: 26 Jun 20221 Jul 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13335 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference4th International Conference on HCI in Mobility, Transport, and Automotive Systems, MobiTAS 2022 Held as Part of the 24th HCI International Conference, HCII 2022
CityVirtual, Online
Period26/06/221/07/22

Keywords

  • Areas of interest
  • Automotive
  • Eye tracking
  • Fleet data

Fingerprint

Dive into the research topics of 'Creating Geopositioned 3D Areas of Interest from Fleet Gaze Data'. Together they form a unique fingerprint.

Cite this