Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

EDGAR: An Autonomous Driving Research Platform -- From Feature Development to Real-World Application

Phillip Karle, Tobias Betz, Marcin Bosk, Felix Fent, Nils Gehrke, Maximilian Geisslinger, Luis Gressenbuch, Philipp Hafemann, Sebastian Huber, Maximilian Hübner, Sebastian Huch, Gemb Kaljavesi, Tobias Kerbl, Dominik Kulmer, Tobias Mascetta, Sebastian Maierhofer, Florian Pfab, Filip Rezabek, Esteban Rivera, Simon Sagmeister, Leander Seidlitz, Florian Sauerbeck, Ilir Tahiraj, Rainer Trauth, Nico Uhlemann, Gerald Würsching, Baha Zarrouki, Matthias Althoff, Johannes Betz, Klaus Bengler, Georg Carle, Frank Diermeyer, Jörg Ott, Markus Lienkamp

TL;DR

The paper presents EDGAR, an autonomous vehicle research platform and its open-source digital twin, designed to close the gap between isolated feature development and full-stack, real-world validation. It details a German-vehicle-based hardware stack, multi-modal sensing, high-performance computing, and a Cyber-Physical Environments workflow that links simulation, HiL, and real-world testing via a common development pipeline. The digital twin encompasses vehicle dynamics, sensor replication, and network replication, enabling repeatable, end-to-end testing and synthetic data generation, while exposing clear gaps between virtual and real-world behavior, particularly in sensor synchronization and calibration. Collectively, EDGAR provides a scalable, open framework for multi-team collaboration, reproducible evaluation, and continuous improvement of autonomous driving software and systems.

Abstract

While current research and development of autonomous driving primarily focuses on developing new features and algorithms, the transfer from isolated software components into an entire software stack has been covered sparsely. Besides that, due to the complexity of autonomous software stacks and public road traffic, the optimal validation of entire stacks is an open research problem. Our paper targets these two aspects. We present our autonomous research vehicle EDGAR and its digital twin, a detailed virtual duplication of the vehicle. While the vehicle's setup is closely related to the state of the art, its virtual duplication is a valuable contribution as it is crucial for a consistent validation process from simulation to real-world tests. In addition, different development teams can work with the same model, making integration and testing of the software stacks much easier, significantly accelerating the development process. The real and virtual vehicles are embedded in a comprehensive development environment, which is also introduced. All parameters of the digital twin are provided open-source at https://github.com/TUMFTM/edgar_digital_twin.

EDGAR: An Autonomous Driving Research Platform -- From Feature Development to Real-World Application

TL;DR

The paper presents EDGAR, an autonomous vehicle research platform and its open-source digital twin, designed to close the gap between isolated feature development and full-stack, real-world validation. It details a German-vehicle-based hardware stack, multi-modal sensing, high-performance computing, and a Cyber-Physical Environments workflow that links simulation, HiL, and real-world testing via a common development pipeline. The digital twin encompasses vehicle dynamics, sensor replication, and network replication, enabling repeatable, end-to-end testing and synthetic data generation, while exposing clear gaps between virtual and real-world behavior, particularly in sensor synchronization and calibration. Collectively, EDGAR provides a scalable, open framework for multi-team collaboration, reproducible evaluation, and continuous improvement of autonomous driving software and systems.

Abstract

While current research and development of autonomous driving primarily focuses on developing new features and algorithms, the transfer from isolated software components into an entire software stack has been covered sparsely. Besides that, due to the complexity of autonomous software stacks and public road traffic, the optimal validation of entire stacks is an open research problem. Our paper targets these two aspects. We present our autonomous research vehicle EDGAR and its digital twin, a detailed virtual duplication of the vehicle. While the vehicle's setup is closely related to the state of the art, its virtual duplication is a valuable contribution as it is crucial for a consistent validation process from simulation to real-world tests. In addition, different development teams can work with the same model, making integration and testing of the software stacks much easier, significantly accelerating the development process. The real and virtual vehicles are embedded in a comprehensive development environment, which is also introduced. All parameters of the digital twin are provided open-source at https://github.com/TUMFTM/edgar_digital_twin.
Paper Structure (32 sections, 1 equation, 17 figures, 8 tables)

This paper contains 32 sections, 1 equation, 17 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Autonomous Vehicle Setup
  4. Vehicle
  5. Sensors
  6. Camera
  7. LiDAR
  8. RADAR
  9. GPS-IMU
  10. Microphones
  11. Computer and Network
  12. High-Performance Computer
  13. Network switch
  14. PTP Clock Synchronization
  15. External Communication
  16. ...and 17 more sections

Figures (17)

  • Figure 1: EDGAR: The research vehicle of the Technical University of Munich.
  • Figure 2: Milestones of AV research vehicles: VaMoRs Maurer1995, Stanley Thrun2006, Boss Urmson2008, MB S-Class Ziegler2014
  • Figure 3: Front roof area with LiDAR and Camera sensors (MR: Mid-Range, LR: Long-Range, SR: Short-Range).
  • Figure 4: Field of view of the perception sensors.
  • Figure 5: Unity environment to simulate point clouds for different LiDAR configurations.
  • ...and 12 more figures