Plenary Speakers
Prof Dr. Christoph Stiller
Karlsruhe Institute of Technology, KIT, Germany
Title: Technology and Safety for the Era of Automated Vehicles
Abstract:
The impending integration of self-driving and cooperative cars into our road traffic landscape heralds a transformative era of mobility. This presentation delves into the technology and safety implications of automated vehicles, spotlighting the pivotal role of video, lidar, and radar sensors in capturing a spectrum of multimodal information from the vehicle’s surroundings. High-definition digital maps further enrich this data, enabling the vehicle to construct a comprehensive model of the real world. Through situational awareness, the vehicle predicts potential future scenarios, laying the groundwork for independent and secure planning and execution of movements to ensure a safe and comfortable traffic flow.
Central to these advancements are the core methodologies rooted in artificial intelligence -perception, scene understanding, and motion planning. A critical aspect explored in this talk is the profound impact of training data on the performance of these methodologies. Additionally, emphasis is placed on safety architectures of the vehicle and its integration into an automation eco-system. Last, not least, the talk sketches field evaluation methodologies of safety promises, as another crucial element in ensuring the reliability and security of automated systems.
Real-world experiments featuring experimental vehicles navigating both non-public roads and natural operations in public traffic will be showcased. The insights garnered from these experiments provide a foundation for the review of feasible solutions and valuable lessons learned, culminating in the identification of open research questions that beckon further exploration in the realm of automated vehicle safety.
Biography:
Christoph Stiller studied electrical engineering at the RWTH Aachen University and at the Norwegian Institute of Technology in Trondheim, Norway. After graduating, he worked as a research assistant at the Institute of Electrical Communication Engineering at RWTH Aachen University, where he received his PhD in 1994. From 1994 to 1995 he was a post-doctoral fellow at INRS-Telecommunications in Montreal, Canada. From 1995, he worked in advance development at Robert Bosch GmbH in Hildesheim in the field of image processing systems. Since April 2001, he has been full professor of the Institute of Measurement and Control Engineering at the Karlsruhe Institute of Technology (KIT).
Christoph Stiller served as president of the IEEE Intelligent Transportation Systems Society (2012-2013). He served as Editor-in-Chief and as associate editor of the IEEE Intelligent Transportation Systems Magazine. His autonomous vehicle AnnieWAY was finalist in the 2007 Darpa Urban Challenge and winner and second winner of the 2011 and 2016 Grand Cooperative Driving Challenge, respectively.
Abstract:
Autonomous vehicle (AV) technology has made significant strides in predicting and avoiding potential collisions. But what happens when avoidance isn’t successful? While the AV stack excels at forecasting risks, it lacks definitive confirmation mechanisms to verify whether a collision was truly avoided — yet such confirmation is essential, as every collision demands an appropriate response.
This gap highlights the critical need for robust impact detection systems that not only enhance vehicle safety but also ensure legal compliance and adherence to the AV industry’s strict reporting requirements.
While experimental contact sensors exist, installing them in all relevant locations on an AV is often impractical. Instead, collision detection should also leverage other sensor data and system outputs, transforming them into reliable impact indicators.
The challenge? Achieving over 99% recall for certain collision types — especially those involving VRUs — while maintaining an exceptionally low false positive rate of no more than one per 100,000 miles, given the severe responses triggered by collision alerts. These challenges are further compounded by the scarcity of high-quality data needed to develop and validate detection systems at scale.
In this keynote, we will explore the technical and operational hurdles of impact detection in AVs and discuss strategies to address data limitations while balancing accuracy and reliability. As the industry moves toward full autonomy, impact detection is no longer a secondary concern — it is a fundamental pillar of AV safety.
Biography:
David Pfeiffer, based in San Mateo, CA, US, is currently a Manager Lidar Perception Team at Zoox. David Pfeiffer brings experience from previous roles at Zoox, Daimler AG, Humboldt University of Berlin and DLR.
David Pfeiffer holds a 2008 – 2011 Doctor of Philosophy (Ph.D.) in Computer Science: Machine Vision for automated vehicles @ Humboldt University of Berlin. With a robust skill set that includes Computer Vision, Image Processing, Autonomous Vehicles, Algorithms, Machine Learning and more.
PD Dr. Victor Pankratius
Robert Bosch GmbH, Germany
Title: Foundation Models in Intelligent Vehicles
Abstract:
GPT technology and foundation models are revolutionizing the landscape of intelligent vehicles, enabling unprecedented advancements in perception, driver assistance, and human-vehicle interaction. This keynote discusses technical progress and challenges in these areas and shows how multimodal data from video, radar, and other sensors can be leveraged to solve long-standing difficult problems.
Looking forward, the keynote outlines how this technology can be further enhanced for deployment into embedded vehicle environments and how ecosystems evolving around foundation models will enable highly adaptive mobility solutions.
Biography:
Victor Pankratius is a Vice President for AI & Data in Bosch’s Cross-Domain Computing Solutions area where he currently advances foundation models for automotive systems, among others. He is an experienced leader with MIT and NASA research background in AI, data science, software engineering, and parallel computing. Victor also served as the global head of software at Bosch Sensortec breaking new ground in AI in mobile devices and wearables.
Prior to Bosch, Victor has led a data science group at MIT and served as a principal investigator in NASA’s prestigious Advanced Information Systems Technology program. Victor earned a Habilitation degree in Computer Science from the Karlsruhe Institute of Technology, Germany, and a doctorate (Dr.rer.pol.) from the University of Karlsruhe’s business school.