Extracting Scenarios from a large amount of real-world driving data increases the value of customer’s test drives. Our Scenario Extractor extracts, identifies, and produces meaningful scenarios from real driver datasets in the most accurate way. Supported by our cutting-edge AI technology, our product facilitates a quick turnaround through a fully automated pipeline.
Scenario Extractor Flow
NavInfo Europe runs an automated scenario extraction process to accurately reconstruct static road environments and dynamic traffic participants by using the dataset from the customer, the information on the required scenarios, and the required test system. The digital scenario layer definition is compliant with the European PEGASUS.
This activity aims to analyze the context and actor behaviors, and automatically identify critical scenarios by classifying event types with relevant attributes and annotations, and applying complexity and risk scoring assessment. Besides, it also discovers unknown instances with the high-risk level and expands our event type classification.
Each extracted scenario is associated with a risk score, which can help clients select the scenarios that fit to validate their Autonomous Driving System for the critical edge cases. It also helps filter a large amount of data and reduce the size of the relevant scenario dataset to eliminate redundant and non-critical scenarios based on clients’ needs.
The Scenario Extractor follows the ASAM standard, ensuring that scenarios can be exported in OpenDrive and OpenScenario to be used in any simulation environment.
Automatic detection of critical traffic events
Accurate replication thanks to robust AI
Assess scenario based on risk scoring system
Ability to support wide-range of sensor setup