Nicole Beringer

Cloud-based validation of automated driving


Dr. Nicole Beringer
Head of Partner Management Germany, Austria, Switzerlandv Elektrobit Automotive GmbH
Am Wolfsmantel 46, 91058 Erlangen, Germany
Phone: +49 (9131) 7701 6226
Mobile: +49-172-4362953
Fax: +49 9131 7701-6333

Ludwig-Maximilians-Universität München
PhD (Dr. phil.) in Phonetics
(automatic speech recognition)
University of Stuttgart
Diploma in Computational Linguistics (Dipl. Ling.)

Work history
Elektrobit Automotive GmbH
Head of Partner Management Germany, Austria, Switzerland
Program Manager Highly Automated Driving, Innovations department
Program Manager Sensor Cloud
cross-domain Project Manager EB Dirigo;
project lead Connected Car Infrastructure, eSecurity,
process consulting,
product management,
supplier management,
SCRUM Product Owner,
SCRUM Master,
workshop moderations,
functional safety,
software development (speech processing),
databases (MySQL, Oracle), APEX applications
Istituto Dalle Molle di Studi sull’Intelligenza Artificiale Lugano
Postdoc Neural Networks, automatic speech recognition
Sympalog Voice Solutions GmbH
Freelancing: pronunciation lexicon French, software development for TextToSpeech French
Conti Temic microelectronic GmbH
Freelancing: pronunciation lexicon French, software development for TextToSpeech French; Technical
translation of user manuals (French)
Ludwig-Maximilians-Universität München
University employee for VERBMOBIL, SmartKom; part project leads with line responsibility for automatic segmentation, End-to-End Evaluation of the SmartKom Dialog system
Freelancing: pronunciation lexicon French, software development for TextToSpeech French

Cloud-based validation of automated driving

Automated vehicles must prove to be reliable in all kinds of complex driving scenarios. This constitutes a major challenge because of the growing complexity of more advanced systems, the need for faster development cycles, and strict legal safety requirements. Real test drive data play a crucial role in the validation task but are not enough on their own. Although there is no commonly agreed number within the industry, numbers for successful validation of a certain function or system are ranging up to billions of high-quality test kilometers. Therefore, we introduce an approach to a comprehensive toolchain to overcome the various challenges ahead, migrating a huge part of the required test kilometers into the virtual world, creating driving scenarios with simulation software. By leveraging this approach with a cloud-computing platform, it becomes possible to easily scale to the power of thousands of high-performance computers, running parallel tests and thus significantly reducing development time. The underlying platform for this toolchain covers data from both real driving test drive and from simulation, creating the basis to cover all required use cases for validation. Furthermore, automation is brought to data administration, enhancement, and labeling, facilitating the handling of the huge amount of test kilometers and petabytes of simulation data. Thanks to the cloud-based approach, test cases can be executed around the world by different developers simultaneously, scaling up to meet critical peak demands. As the overall approach is designed openly, it will be possible to create a truly comprehensive solution together with partners, answering the specific needs of car makers and suppliers.