面向自动驾驶的可解释人工智能技术研究

打开文本图片集
中图分类号:U471 收稿日期:2025-04-22 DOI: 10.19999/j.cnki.1004-0226.2025.06.002
Abstract:Artificialiteligencetechnologyrepresentedbydeplearingpromotes therapiddevelopmentoftheautonomous drivingindustry,ndteblackboxproblemofrtificialintellgencenetworksiscomiganimportatobstacletotepopularzatiofu tonomousvehicles.ExplainableArtfcialIntelligence(XAI)technologyisanefectivemethodtosolvethebackboxproblem.Facing theimmaturityofXItechnology,thispaperproposesthedefinitionofXAIforautomaticdriving,andcomprehensivelyanalyesthe relevantcasestudy.FurtherthispaperproposesaframeworkofXAIechnologyforutonomousdriving,andofersanin-depthelucidationofitskeyompoents.ThispaprsystematicallalyesXAItechologiesfoutoomousdriving,isussstetallenges,andlooksforwardtothefutureresearchdirection,hopingtoprovideguidanceandreferenceforthefuturedevelopmentofau tonomous driving and XAI.
Key words:Autonomous driving;ExplainableArtificial Inteligence(XAI);Technical framework;Black box
1前言
以深度学习为代表的人工智能技术是自动驾驶领域的核心技术。(剩余8481字)