Scenario Extraction for Self-Driving Vehicles - Real World Applications

Explore a 26-minute conference talk on scenario extraction for self-driving vehicles presented by Fabian Peng Kärrholm, Senior Data Scientist at Volvo Cars. Delve into the critical process of collecting and analyzing real-world scenarios for validating autonomous driving functions. Learn about methods for evaluating time-series data from vehicles, classifying it into states, events, and scenarios, and using these findings for simulations, risk analysis, and traffic behavior assessment. Gain insights into various data analysis and structuring techniques, as well as approaches to draw meaningful conclusions about traffic behavior. Discover the speaker's diverse background, including his PhD in Combustion & Thermodynamics and his current focus on scenario extraction, identification, and classification for traffic analysis in the Autonomous Drive department at Volvo Cars. This talk was recorded at the 2024 GAIA Conference on March 27 at Svenska Mässan in Gothenburg, Sweden.