10 Years Of Ai In Autonomous Driving: Lessons Learned From Ai Adoption For Sae L3 And L4 Automated Driving Functions, With An Outlook On Adaptation Needs

With the first commercialization approval in San Francisco and the inaugural SAE L3 traffic jam pilots hitting European streets, the AV/ADAS industry is entering a new phase — the technology's small-scale maturity has been proven. However, for ADAS/AD technology, and AI in particular, to scale successfully, the integrity and confidence in the system must improve. The deployment of these safety-critical AI systems on public roads, as seen in China, the US, and Europe, has shifted responsibility from the driver to the system, necessitating high standards in quality service, robustness, tool confidence, and safety. AI features in autonomous driving in two key ways: it is adapted for decision-making within the self-driving system technology stack, and it plays a part in the development toolchain for simulation and virtual validation. Debate over the adequacy of the transformer or CNN model approach and the appropriate size of data sets continues. While integrating AI has its merits, challenges such as achieving compliance, promoting trust through explainable AI, and reaching automotive readiness persist. This session will foster discussions on these points, validate procedures, strategize compliance, and explore potential reasons for the current industry gaps. Moreover, it will offer key insights that can be applied across various aspects of autonomous effectivenessTakeaways from the Talk:

​

• Progress and Challenges of AI: Learn about the evolution and the current of AI in Autonomous Vehicles (AV)/Advanced Driver Assistance Systems (ADAS).

• Responsibility Shift in AI Systems: Discuss the importance of quality, robustness, and safety as the liability shifts from the driver to the system.

• AI's Role in Decision-Making and Development: Understand AI's influence on autonomous driving technology decisions and simulation development tools.

• Overcoming AI Integration Obstacles: Gain insights into compliance, trust promotion via explainable AI, and achieving readiness in the automotive industry.