大模型在智能网联汽车上的关键技术研究与应用  

Research and Application of Key Technologies of Large Models in Intelligent Connected Vehicles

作　　者：赵星洋 黄爱林 ZHAO Xingyang;HUANG Ailin(School of Control and Computer Engineering,North China Electric Power University,Beijing 102200,China;Shenzhen Hangsheng Electronics Company Limited,Shenzhen 518000,China)

机构地区：[1]华北电力大学控制与计算机工程学院,北京102200 [2]深圳市航盛电子股份有限公司,广东深圳518000

出　　处：《汽车实用技术》2025年第8期29-36,共8页Automobile Applied Technology

摘　　要：随着智能网联汽车的准入和渗透的不断加深,通过大模型技术推动产业发展和应用落地逐渐成为了业内共识。文章对大语言模型、视觉语言大模型、多模态大模型的技术架构、算法和原理进行了系统研究,并对其在智驾域、座舱域、网联域中各场景的具体应用进行了深入分析。通过对当前市场主流车企车型及技术案例的研究,详细探讨了大模型在多域智能上的实际应用现状,揭示了大模型在推动智能网联汽车技术进步中的关键作用。最后,对目前大模型技术所面临的挑战进行了分析,并对今后的发展提出了有针对性的建议。文章对大模型在智能网联汽车领域的技术研究具有重要参考价值。With the continuous deepening of the access and penetration of intelligent connected vehicles,promoting industrial development and application implementation through large-model technology has gradually become the consensus within the industry.This article conducts a systematic study on the technical architectures,algorithms,and principles of large language models,visual-language large models,and multimodal large models,and conducts an in-depth analysis of their specific applications in various scenarios in the intelligent driving domain,cockpit domain,and networking domain.Through the study of the models of mainstream automobile enterprises and technical cases in the current market,it discusses in detail the actual application status of large models in multi-domain intelligence and reveals the key role of large models in promoting the technological progress of intelligent connected vehicles.Finally,it analyzes the challenges currently faced by large-model technology and puts forward targeted suggestions for future development.This article has important reference value for the technical research of large models in the field of intelligent connected vehicles.

关 键 词：大模型 智能驾驶 智能座舱 智能网联 算法架构 市场渗透 

分 类 号：TP3-05[自动化与计算机技术—计算机科学与技术] U461[机械工程—车辆工程]