leveLXData by fka

⚠️ Calling all researchers! Check this #CallforPapers for the Workshop on Interaction-driven #BehaviorPrediction and Planning for #AutonomousVehicles at #IV2026 in Detroit! Base your contributions on reliable and high-quality data: Our public, drone-based trajectory datasets, which contain information on hundreds of thousands of road users, are ideal for developing and testing your ideas. Each dataset includes comprehensive Lanelet2 and OpenDRIVE HD maps, along with Python tools designed to help you kickstart your projects effortlessly. 🟠Highway Datasets: - highD dataset (https://lnkd.in/dtP7TYX3): Focuses on straight free flowing highway traffic at different speed limits. - exiD dataset (https://lnkd.in/d8uTtvNc): Captures interactions on highways with on- and off-ramps for more complex scenarios. 🟠 Urban Datasets - inD dataset (https://lnkd.in/dTwpME2n): Unsignalized intersections with varying speed limits, including lots of vehicle VRU interactions. - rounD dataset (https://lnkd.in/diuucpGw): Roundabouts of different sizes and speeds, with VRU data included. - uniD dataset (https://lnkd.in/dVgxpMjC): Focused on shared urban spaces with the highest number of VRUs. ➡️ Learn more and submit your work: https://lnkd.in/d6Zh45S2 #IEEE #IV2026 #AutomatedMobility #AutomatedVehicle #BehaviorPrediction #BehaviorPlanning

 New Research Spotlight: LC-LLM & the Power of High-Quality Traffic Data! We're excited to share a publication showcasing the application of Large Language Models (#LLM) for enhanced #lanechange #prediction in automated driving! 🚗💨 Mingxing Peng, et al. from the HKUST Guangzhou have developed LC-LLM, a model that not only forecasts lane change intentions and trajectories but also provides explainable reasoning behind its predictions. This is a huge step forward for building trust and safety in autonomous systems! A key ingredient in this success is high-quality traffic data! The team leveraged data from #leveLXData, specifically utilizing our rich data to train and validate their model. This demonstrates the critical role that accurate, real-world data plays in advancing automated driving. The study highlights how leveraging LLMs combined with comprehensive datasets lead to: ✅ More accurate prediction of vehicle behavior ✅ Improved interpretability of decision-making processes ✅ Enhanced safety and reliability of autonomous systems 🔗 Find the links to the paper, the #highD #dataset, and the github project in the comments below! #AutonomousDriving #AI #MachineLearning #TrafficData #Innovation #Safety #Research #ADAS #IntelligentVehicles

🎄The LXD team says thank you and sends the best wishes to our partners, customers and friends! We are looking forward to keep on supporting #safe, #efficient and #comfortable #mobility with great innovations in 2026! Stay tuned, we have some fancy new tech and data coming up 😉

🎂 Happy Birthday, #highD! At IEEE ITSC 2018 highD, the Highway Drone Dataset, was officially published and set a new standard for precise naturalistic driving data. The novel methodology enabled not only efficient #scenario data collection, but delivers valuable data for so many applications. Seven years, thousands of downloads and about 2000 paper citations later, its relevance is still unbroken. Get your copy: www.highd-dataset.com

This is a great chance to gain some first-hand experience how to get the most out of your scenario data. BTW: Our #inD dataset can be explored immediately at www.scenario.center

🚗🚛 Visit at the Final Event of the Hi-Drive Project It was really exciting to see all the great results of the work from the past years. We congratulate all partners of the Hi-Drive consortium for the successful conclusion and great presentation of their achievements in yesterday's final event! With their collaborative spirit and innovative ideas, researchers, industry leaders, technology providers, universities, and regulatory bodies have tackled some of the biggest challenges in automated vehicle deployment, focusing on expanding #OperationalDesignDomains, enhancing safety, and reducing reliance on human intervention. We are happy, that we could support the project with trajectory datasets from #leveLXData, especially for implementing and testing #EdgeCase Detection approaches. Saeed Rahmani. Jonas Bärgman, Henri Sintonen, Marcel Sonntag Be sure to check out the Hi-Drive Driving Scenario and Edge Case Database (DSDB), which has been set up by our colleagues from Institut für Kraftfahrzeuge (ika) - RWTH Aachen University and compiles interesting data insights from the project: https://lnkd.in/egD7GBvY Now we are looking forward to what's coming up next! #HiDriveFinalEvent #AutomatedDriving #Collaboration #EdgeCaseDetection #ScenarioAnalysis #Innovation #FutureOfMobility

⚡Your self-driving stack hasn't seen it all. Ensuring the #safety and reliability of automated vehicle in all driving #scenarios, particularly edge cases and critical situations, is important. These rare but significant events can test the limits of an AV's decision-making capabilities and are essential for comprehensive testing and #validation. These #edgecases occur infrequently or under unusual conditions, and are therefore expensive to capture. Leveraging leveLXData's high-precision drone-based trajectory data from challenging environments worldwide, helps to lower cost and still achieve wide scenario #coverage. Missing an off-ramp on a motorway can be an example for such edge cases, as shown in the video below, which was filmed on a motorway in Arizona. 👉 Get in touch with Thomas Keutgens and Christoph Klas to get a demo! #CriticalScenarios #AutonomousDriving #SafetyFirst #leveLXData #BEVdata #ADAS 

🚗🚀 Join the Conversation on Scaling Autonomous Driving! Meet us at the "Tagung Automatisiertes Fahren" by TÜV SÜD and Technische Universität München. This conference addresses a range of topics from technical requirements to commercial applications and business models for autonomous driving. Discussions will focus on both modular and end-to-end approaches, emphasizing the importance of simulation in enhancing reproducibility and transparency during validation processes. Meet Thomas Keutgens to discuss how our cutting-edge technology can help scale autonomous driving systems throughout development and validation phases. Whether you're a developer, researcher, or business strategist, this is your opportunity to get fresh ideas on how to scale autonomous driving. 👉 Let’s connect! #AutonomousDriving #leveLXData #TagungAutomatisiertesFahren #TUVSüd #TUMunich 

This is a great example of how our #trafficdata can provide insights into #AV operations by different providers, like Waymo, Zoox and Tesla. Can't wait to catch a MOIA vehicle in #Austin as well. Thanks to Michael Woon and the Retrospect team for the productive cooperation!

Meet our colleagues in Tempe, AZ at #SAE #ATS2025. In their traditional Breakout Session, they will discuss the latest developments and ideas in #SafetyAssurance of automated driving and how high-quality #trafficdata supports it. With great contributions of Waymo, Zoox, Torc Robotics, Aurora, Hiroki Nakamura from JARI amongst others. Adrian will also present some interesting learnings and preliminary results from the European Hi-Drive project in the Wednesday plenary session. Contact Lutz Eckstein and Adrian Zlocki, if you are at #ATS2025 or in the Phoenix region. They are happy to chat! #automateddriving #safety #ADAS #AD #DroneData #ScenarioData