Logo des Repositoriums
  • English
  • Deutsch
Anmelden
Keine TU-ID? Klicken Sie hier für mehr Informationen.
  1. Startseite
  2. Publikationen
  3. Publikationen von Externen
  4. Erstveröffentlichungen (extern)
  5. EDAG Group - Edlight & Homogenium
 
  • Details
2025
Erstveröffentlichung
Konferenzveröffentlichung
Verlagsversion

EDAG Group - Edlight & Homogenium

File(s)
Download
Hauptpublikation
Buthmann_EDAG Group - Edlight & Homogenium.pdf
CC BY 4.0 International
Format: Adobe PDF
Size: 130.14 KB
TUDa URI
tuda/14169
URN
urn:nbn:de:tuda-tuprints-308837
DOI
10.26083/tuprints-00030883
Autor:innen
Buthmann, Jannes Alexander
Metzger, Julian
Kurzbeschreibung (Abstract)

Contemporary automotive lighting systems are becoming increasingly sophisticated, incorporating advanced functionalities such as Adaptive Driving Beam (ADB), Adaptive Front Lighting Systems (AFS), and dynamic animations. This growing complexity poses significant challenges in development, integration, and especially in testing. Conventional manual validation procedures are characterised by their protracted nature, susceptibility to errors, and inadequate scalability. In order to address these challenges, EDAG has developed two innovative tools: EDLIGHT and HOMOGENIUM. EDLIGHT is an AI-driven solution for the automated validation of light functions. The employment of deep learning models, such as ResNet18 and YOLOv5, enables a substantial reduction in the time and manual effort required for testing. This is achieved by leveraging these models to perform end-to-end validation of beam patterns and lighting modes. EDLIGHT is a software that automates the entire testing workflow for automotive derivatives, encompassing all aspects from requirement definition to result reporting. It has been empirically demonstrated that the utilisation of EDLIGHT results in a significant reduction in test time, from 120 hours to 30 hours, thereby achieving a 75% efficiency gain. The primary benefits of this approach include the assurance of consistency, the provision of safety for engineers (by reducing light exposure), and the capacity for future adaptation to accommodate additional lighting features. In contrast, HOMOGENIUM offers a precise and objective evaluation of light homogeneity and luminance distribution. The methodology employed is a stepwise process that involves the filtration of images, binarisation, labelling, and morphological smoothing, with the objective of producing detailed gradient and homogeneity maps. The outputs are then evaluated against OEM specifications using standard, complex, or custom mathematical methods. The result is a fully automated, repeatable, and highly accurate analysis that reduces manual simulation and evaluation time from 4--8 hours to under 1 hour. The collaboration between EDLIGHT and HOMOGENIUM signifies a substantial progression in the validation of automotive lighting systems. The utilisation of these tools has been demonstrated to facilitate accelerated development cycles, enhance test coverage, and ensure superior product quality. It is evident that these tools offer a scalable and future-proof framework for Original Equipment Manufacturers (OEMs) and suppliers.

Sprache
Englisch
Herausgeber:innen
Khanh, Tran Quoc ORCID 0000-0003-1828-2459
DDC
600 Technik, Medizin, angewandte Wissenschaften > 620 Ingenieurwissenschaften und Maschinenbau
Institution
Universitäts- und Landesbibliothek Darmstadt
Ort
Darmstadt
Veranstaltungstitel
16. International Symposium on Automotive Lighting (ISAL)
Veranstaltungsort
Darmstadt
Startdatum der Veranstaltung
22.09.2025
Enddatum der Veranstaltung
24.09.2025
Buchtitel
Proceedings of the 16th International Symposium on Automotive Lighting 2025
Titel der Reihe
Darmstädter Lichttechnik
Bandnummer der Reihe
21
PPN
53392166X

  • TUprints Leitlinien
  • Cookie-Einstellungen
  • Impressum
  • Datenschutzbestimmungen
  • Webseitenanalyse
Diese Webseite wird von der Universitäts- und Landesbibliothek Darmstadt (ULB) betrieben.