we.CONECT spoke to Dr. Marco Jähnisch, Head of ADAS & AD Software Product Lines, Situation Interpretation & Vehicle Behavior at CARIAD, who is a session moderator at the 6th Auto.AI Europe.
we.CONECT: Hi, Marco, tell us a few words about your role at CARIAD. What makes you especially passionate about your job?
Marco Jähnisch: My Team and I develop ADAS/AD Software focusing on software product lines, Situation Interpretation, and Vehicle Behavior. We locate in different locations like Berlin, Stuttgart, and Wolfsburg, but we are united in the passion for software and our products.
we.CONECT: As head of ADAS & AD, what do you think are the most promising technologies to meet the challenges faced by ADAS and AD engineers?
Marco Jähnisch: I think the capability to manage and analyze big data and deploy software and datasets in short cycles to the fleet will be an excellent foundation to face the challenges in the ADAS and AD field.
we.CONECT: Your career in software development spans two decades, what is your vision for software in cars?
Marco Jähnisch: The software will be more and more hardware independent. Because of the higher reliability of the connectivity, it will be easier to shift between the onboard computer unit and the backend cluster compute. This approach gives you more flexibility to scale your computing tasks.
we.CONECT: You will be moderating the Software/Hardware Café at Auto.AI, what is it about the interactive format that you like the most?
Marco Jähnisch: I like to contact the participants and discuss ideas and other perspectives on questions.
we.CONECT: What are the biggest challenges in data processing for automated driving, in your opinion?
Marco Jähnisch: I see at least two challenges to overcome. At first, you must be able to handle and manage the amount of data in the fleet accessible. Second, you need suitable data analysis to determine which data to answer the question and which information may be missing (e.g., corner cases).
we.CONECT: You are a returning participant at our event. What are your expectations from this year’s Auto.AI?
Marco Jähnisch: I’m looking forward to getting in touch with the participants and to having fruitful and inspiring talks.
we.CONECT: Thank you for your time, Marco, we are looking forward to your session at the conference!
Marco Jähnisch will be moderating a SW/HW Café and exploring the topic of Data processing and AI – software architectures and hardware challenges on September 30 at 11:40.
AI Approaches could be used for a wide field of applications and achieved enormous benefits in comparison to standard approaches. For a suitable application of such AI approaches it is necessary to understand some differences between standard approaches and Ai approaches from the perspective of software architecture and hardware challenges.
- What are the most challenging topics to bringing AI approaches in a cost-efficient way into serial software development?
- Are Ai-based approaches always the best solution or have a hybrid approach sometimes a befit?
- What kind of compute unit is needed?
- Which challenges do we have to face for AI approaches in software architecture?
- Which methods are suitable for the validation and verification of Ai approaches?
Auto.AI Europe is the leading technical event on deep learning for SAE Level 4 & 5 automated vehicles that brings together more than 300 top-industry machine learning, neural networks, and perception experts and decision-makers. Join now and discuss self-supervised and behavioral learning concepts, scalable machine and reinforcement learning approaches, and benchmarking perception and computer vision systems for ADs with your peers from the automotive AI community.