Waymo’s Post

Our paper, "Exploring the Potential of World Models for Anomaly Detection in Autonomous Driving", has been accepted at the IEEE Symposium Series on Computational Intelligence 2023. We showed that world models can be transferred from model-based reinforcement learning to the task of anomaly detection in autonomous driving. World models can be used to implement existing ideas from current approaches in a unified anomaly detection framework. The preprint is available on arXiv: https://lnkd.in/ejFyJQCB

Dear Friends, Together with my colleagues Dan Doru MICU and Bogdan Ovidiu VARGA released a new Special Issue with the objective of collecting virtual models for autonomous driving systems. "Virtual Models for Autonomous Driving Systems" intends to compile autonomous driving simulation methods and scenarios into a single COMPEDIUM to provide scientists and researchers with a clear methodology for validating autonomous vehicles. We invite all interested authors to publish open access papers in Applied Sciences MDPI on autonomous driving virtual systems and algorithms, including virtual environments, routes, infrastructure, and traffic models, simulation methodology and scenarios for connected vehicles, and collaborative driving integrated into intelligent transportation systems. Link: https://lnkd.in/emuRSiey

Autonomous driving has shown great potential to reform modern transportation, yet its reliability and safety have drawn a lot of attention and concerns. In our latest research seminar, we were joined by xi zheng, Director of Intelligent Systems Research Group at Macquarie University, and CCSRI member and RMIT University Lecturer, Xiaoyu (Shaw) XIA, for an in-depth discussion on the topic, including insights into Dr. Zheng's research and the challenges of software testing. If you missed it, you can watch the seminar, Towards Robust Autonomous Driving Systems, on our YouTube channel or linked below. https://lnkd.in/gtqJxNZh #autonomousdriving #autonomoussystems #autonomousvehicles

A new book is getting published on Autonomous vehicles focusing on the multifaceted technological and societal perspectives. It has been curated by Dr. Ishwar Sethi from Oakland University. I have contributed a chapter on state of simulation and modeling, talking about: - the various aspects of scene, scenario and situation, - the different ways to structure simulation scenarios in different countries, - sensor simulation and - problems and oppurtunities. The link to the book: https://lnkd.in/gQ7K-vQr

New to autonomy and want to get up to speed? This session from #GTC23 covers how autonomous vehicles work. From core technologies like GPU compute, Omniverse and Generative AI to the latest breakthroughs enabling safe autonomous driving. Watch now.

In the most recent paper published by Zenseact PhD students we present a versatile NeRF-based simulator for testing autonomous driving (AD) software systems, designed with a focus on sensor-realistic closed-loop evaluation and the creation of safety-critical scenarios. The evaluation reveals that, while state-of-the-art end-to-end planners excel in nominal driving scenarios in an open-loop setting, they exhibit critical flaws when navigating our safety-critical scenarios in a closed-loop setting. This highlights the need for advancements in the safety and real-world usability of end-to-end planners. By publicly releasing the simulator and scenarios as an evaluation suite, the research community is invited to explore, refine, and validate their AD models in controlled, yet highly configurable and challenging sensor-realistic environments. The great work was done by the talented PhD students William Ljungbergh, Adam Tonderski, Joakim Johnander as well as Holger Caesar, Kalle Åström , Michael Felsberg and Christoffer Petersson. Thanks to WASP – Wallenberg AI, Autonomous Systems and Software Program #towardszerofaster Project Page: https://lnkd.in/dSv7HFYQ Code: https://lnkd.in/d73HcWrb Paper: https://lnkd.in/dTQfg__9

🚗 Excited to share a new research paper on autonomous decision-making! Check out the latest blog post introducing the Driver to Evaluation dataset (D2E) - a comprehensive dataset covering driver states, vehicle conditions, and human evaluation scores. This dataset goes beyond traditional vehicle and environmental data, providing crucial insights into human factors. Dive into the details at https://bit.ly/3xeFkhg to learn more about how D2E is shaping the future of data-driven decision-making in autonomous driving. #AutonomousDriving #DataScience #ResearchPublication

New to autonomy and want to get up to speed? This session from #GTC23 covers how autonomous vehicles work. From core technologies like GPU compute, Omniverse and Generative AI to the latest breakthroughs enabling safe autonomous driving. Watch now.

New to autonomy and want to get up to speed? This session from #GTC23 covers how autonomous vehicles work. From core technologies like GPU compute, Omniverse and Generative AI to the latest breakthroughs enabling safe autonomous driving. Watch now.

New to autonomy and want to get up to speed? This session from #GTC23 covers how autonomous vehicles work. From core technologies like GPU compute, Omniverse and Generative AI to the latest breakthroughs enabling safe autonomous driving. Watch now.