US 12,430,341 B2
Scenario similarity retrieval-based automatic scenario generation system and method
Chang Gue Park, Seoul (KR); Kyoung Won Min, Seongnam-si (KR); Haeng Seon Son, Seongnam-si (KR); Seon Young Lee, Seoul (KR); Young Bo Shim, Seongnam-si (KR); Gi Ho Sung, Seongnam-si (KR); Jin Man Park, Gwangju-si (KR); and Yeong Kwon Choe, Seongnam-si (KR)
Assigned to Korea Electronics Technology Institute, Seongnam-si (KR)
Filed by Korea Electronics Technology Institute, Seongnam-si (KR)
Filed on Dec. 28, 2022, as Appl. No. 18/089,674.
Claims priority of application No. 10-2022-0166795 (KR), filed on Dec. 2, 2022.
Prior Publication US 2024/0184787 A1, Jun. 6, 2024
Int. Cl. G06F 16/2457 (2019.01); G07C 5/08 (2006.01)
CPC G06F 16/24575 (2019.01) [G07C 5/085 (2013.01)] 9 Claims
OG exemplary drawing
 
1. A processor-implemented method of validating an autonomous driving using scenario retrieval-based automatic scenario, the method comprising:
retrieving scenarios similar to a query scenario from a scenario database (DB) for an autonomous driving test;
filtering only scenarios that meet a selection condition from the retrieved scenarios;
converting components of traffic situations included in the filtered scenarios, by sampling parameters associated with the filtered scenarios, to generate the scenario retrieval-based automatic scenario associated with the query scenario, including merging the filtered scenarios into a map; and
validating the autonomous driving by performing an autonomous driving test using the generated scenario retrieval-based automatic scenario,
wherein the converting comprises converting an abstract scenario into a logical scenario, and concretizing a logical scenario into a concrete scenario by sampling a specific value of the parameters based on probability distribution,
wherein the retrieving comprises:
calculating a similarity by comparing components of the query scenario with components of scenarios stored in the DB;
sorting scenarios with reference to the calculated similarity; and
selecting scenarios that have a similarity greater than or equal to a predetermined threshold, and
wherein the calculating the similarity comprises:
extracting feature values from the scenarios; and
calculating a similarity between the extracted feature values,
wherein the selection condition comprises an operational design domain (ODD), criticality phenomena (CP), and a filtering condition,
wherein the ODD limits an operational range of a system with respect to factors, related to driving, including an operation on a road and the road, and
wherein the filtering condition comprises a scenario abstraction step, an effectiveness validation method, and a position where a scenario is collected.