Join the Team
Become part of the EDGAR team and participate in applied research on the autonomous driving of tomorrow. We offer PhD positions as well as various opportunities for students.
The Idea
A large number of algorithms already exist in the state of the art that focus on the individual subtasks of autonomous driving, such as image recognition or vehicle longitudinal control. The great challenge, however, is still to develop an optimal overall software for the autonomous driving task from the individual modules. The application of the individual functional modules in the overall software results in new dependencies and interlinked effects. In addition, the hard constraints of limited computing capacity in the vehicle and the lowest possible latencies come into play. Our research focus is on this overall software approach with the goal of real applications on the target hardware. This allows us to precisely specify the critical molecules in the software and the intermodal sensitivity, and to derive research topics that relate directly to a specific application-relevant problem.
The application of the developed software under real conditions takes place on a Level-5 capable research vehicle, which is equipped with extensive sensor technology for recording the environment and the own vehicle state, high-performance computers for the application of the software as well as interfaces for the control of the longitudinal and lateral guidance. In addition, the high-performance computers are embedded in an extensive simulation environment, which is used during the whole development phase and for the automated testing of the software functionality in the run-up to real vehicle tests. The research vehicle incl. data center and simulator are funded by a proposal at the German Research Foundation (DFG) (approval according to Art. 91b GG with DFG-number INST 95/1653-1 FUGG).
The Goal
Our goal is always the specific application-driven further development of existing algorithms with regard to the optimal overall software. For this purpose, we look for complex, selected application areas (operation design domain, ODD), which we use to demonstrate our software functionality and for the final validation of our development work. Our next target ODD, in which we want to demonstrate driving with our development vehicle and the developed software, are urban scenarios in inner-city traffic of Munich as well as hub-to-hub highway driving. Despite the specified target ODD, we always pursue the goal of developing an overall software that is as generic as possible and can be applied to different road users, road geometries, speed ranges, and weather conditions. The extensive sensor technology of the research vehicle will also be used to build a multimodal data set that will be published to support development activities.
The Team
The FTM has the goal of combining the core competencies of various chairs in this project and creating synergies in the development and application of the research content. The following members of the Chair of Automotive Engineering are currently involved in the project:
| Nico Uhlemann Prediction | Trajectory prediction Behavioral models Pedestrian simulation |
| Tobias Betz Compute Platform | Software Orchestration Latency Optimization Vehicle Performance |
| Rainer Trauth Vehicle Performance | Explainable AI Simulation Vehicle Performance |
| Felix Fent Perception | Multi-Modal Perception Perception Sensor Calibration |
| Baha Zarrouki Vehicle- & Motion Control | Project: adaptive MPC for trajectory following Model Predicitive Control Vehicle Dynamics Simulation |
Cyber-Physical Systems, Prof. Althoff
Chair of Ergonomics, Prof. Bengler
TUM Research Groups
The consortium of the DFG research vehicle includes the following partner chairs and institutes at the TUM:
Partner
Within various individual projects, the Chair of Automotive Engineering cooperates with the following partners:
