http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
엣지 인프라 기반 자율주행 기능의 고장 안전성 평가를 위한 오류 주입 시나리오 연구
박종기(Jongki Park),신성근(Seonggeun Shin),예창민(Changmin Ye),우창수(Changsu Woo),이혁기(Hyuckkee Lee) 한국자동차공학회 2022 한국자동차공학회 학술대회 및 전시회 Vol.2022 No.11
Recently, development and research on the technology of autonomous vehicles have been actively conducted in Korea. Among them, self-driving vehicles based on edge computing that can process a large amount of real-time data at close range have the advantage of being more stable and reliable. Unlike cloud computing, which has a centralized structure, edge computing is a technology that processes at a midpoint between networks without relying on a central server and is processed in real-time from close range. It is evaluated as a more stable autonomous driving function due to reduced network latency and increased reliability by processing vast amounts of sensors and various communication data inside the vehicle and directly from V2X communication terminals. To improve the reliability and safety of these edge data, the risk of failure of these features should also be addressed along with the study of edge infrastructure-based autonomous driving features. Therefore, this paper deals with error injection scenarios for failure safety evaluation of such edge infrastructure. The target is Daegu Technopolis, which deals with error injection scenarios to ensure the safety of failures of autonomous driving functions for three types of edge infrastructure unexpected situation detectors, pedestrian detectors, and signal controllers in empirical areas. First, the edge infrastructure environment and functions of the infrastructure are defined, and the malfunction and risk causes of the edge infrastructure are analyzed using the HAZOP technique, and the failure risk scenario of the edge infrastructure is derived through HARA analysis to perform the edge infrastructure error injection test in the simulation environment. Through this procedure, an error injection scenario for failure safety evaluation of edge infrastructure-based autonomous driving functions was studied. Based on the scenarios derived in the future, the safety evaluation method of edge infrastructure-based autonomous driving functions will be studied through error injection tests in a simulation environment.