大模型驱动的智能汽车音效算法与系统研究综述

打开文本图片集

中图分类号：U463.671 文献标识码：A 文章编号：1003-8639（2026）01-0043-05

A Review of Research on Vehicle Sound Effect Algorithms and Systems Driven by Large Language Models

LiuYanfei，ZhangJie （IFLYTEKCo.，Ltd.，Automotive BusinessUnit，Hefei 23oo88，China）

【Abstract】With the rapid development and intelligent evolution of large language models（LLMs）and the automotive industry，sound effect algorithms for vehicleplaya crucial role inenhancing the acoustic experience within thecabin.Inthis paper，a comprehensive reviewofsound efect algorithms，functions，systemarchitectures，and applications for vehicle basedon theLLMs was provided.Firstly，itintroduces thedevelopment historyand current technological statusofthein-vehiclesoundeffects，alongwithfutureapplicationscenariosandtrendsinthis fielddriven byLLMs.Next，itanalyzes the advantages of more than 1O product functions and algorithms from three dimensions： soundreproduction，noise management，and intelligent experience.The paper furtherdiscusseschallenges and near-term solutionsforimplementingthesealgorithms invehicles.Finaly，Large language modelsconductdata-drivenlearningand model training onaudiosystems of commerciallavailable vehicles.This process enabled thedevelopmentofareference framework for enhancing in-vehicle acousticsystemsacrossthree critical dimensions： audiochip，amplifiers，andspeaker. This review aims to serve as areference for research on LLMs based smart car sound efects，promoting widespread adoption in the automotive field.

【Key words】LLM; vehicle sound effect;audio product; sound system

0 引言

中国汽车工业正向电动化、智能化加速发展，汽车已成为第三移动空间，智能座舱作为用户易感知的智能化领域，其音频能力直接影响驾乘体验。（剩余7908字）