1. 자동차용 음성의 탑재율은 83%이 넘고, 고급 음성 기능의 탑재율이 대폭 증가합니다.
2024년 1월-11월, 자동차용 음성 시스템의 탑재수는 1,676만대, 탑재율은 83.3%가 되었습니다. 2023년도와 비교하면 탑재율이 5%포인트 상승했습니다. 에너지 유형별로는 EREV(Extended-Range Electric Vehicle)의 탑재율이 가장 높고, 2024년 1월-11월의 탑재율은 100%에 달했습니다. 이 에너지 유형의 대표적인 모델은 Li Auto L 시리즈, AITO M 시리즈, Deepal S 시리즈 등입니다.
음성 기능은 2024년 연속 대화, 씨앤스피크, 웨이크업 프리 기능의 탑재 수와 탑재율이 크게 증가했습니다.
씨앤스피크 기능에서는 2024년 1월-11월의 누계 탑재수는 466만대, 탑재율은 23%로 2023년 통년에 비해 18%포인트 증가했습니다. 에너지 유형별로는 EREV의 탑재율이 92.1%로 가장 높고 연료차는 7.1%로 가장 낮습니다. 가격대별로는 50만 위안 이상에서는 씨앤스피크 기능의 탑재율이 가장 높고 Zeekr 009, Yangwang U8, NIO ES8 등이 대표적인 모델입니다. 또, 이 가격대의 탑재율 증가폭은 가장 크고, 48%포인트 증가했습니다. 이는 2024년에 자동차 음성 시스템의 지능 수준이 크게 향상됨을 보여주었습니다.
2. 조종석은 더 많은 생태 자원에 액세스하고 음성 어시스턴트는 심층 서비스 기능을 확보합니다.
기본 모델 시대에 "많은 것을 알고 서비스를 제공할 수 있는" 음성 어시스턴트는 다양한 생태계 용도에 대한 액세스에 크게 의존합니다. 예를 들어, 사용자가 '자동차가 꺼질 것 같다', '배고파', '설날에 무엇을 입으면 좋을까'라는 막연한 명령을 내린 경우, 음성 어시스턴트의 응답에는 지도, 지역 생활 서비스, 온라인 정보 등의 용도으로부터의 지원이 필요합니다.
AMAP, iQiyi, Tencent Video, NetEase Cloud Music, QQ Music과 같은 일반적인 용도 외에도 Li Auto는 Xiaohongshu(Little Red Book) 플랫폼의 컨텐츠에 대한 음성 통화를 구현하고 Meituan을 위해 깊은 맞춤형 음성 기술을 시작했습니다. 예를 들어 'Xiaohongshu가 추천하는 설날 복장', 'Xiaohongshu에서 베이징 여행 가이드 찾기', 'Meituan에서 평균 가격 200위안, 평점 4.5초의 광동 요리 레스토랑 찾기' 등을 요청하기 위해 사용자는 Lixiang Tongxue를 시작할 수 있습니다.
3. 파운데이션 모델 애플리케이션은 '명령 상호 작용'에서 '인지 상호 작용'으로 자동차 음성 개발을 가속화합니다.
기존의 명령 상호 작용과는 달리, 기반 모델에 의해 강화된 자동차 음성 시스템은 말하기 이해, 논리적 추론, 지식 Q&A, 회화 작성, 차량 주변 환경 인식에 보다 우수한 능력을 발휘합니다.
예를 들어, XPeng의 XGPT를 탑재한 Xiao P 어시스턴트는 음성 언어 이해, 논리적 추론, 지식 백과 사전, 회화·이야기·동화의 작성, 차량 주위의 물체의 인식 등의 기능을 갖추고 있습니다.
Li Auto의 Mind GPT를 탑재한 Lixiang Tongxue는 Lixiang Tongxue에 「영화의 이름을 잊었습니다. 흑인의 피아니스트가 나오는데, 무슨 영화인지 알겠어?」라고 물어보는 모호한 검색 기능, Lixiang Tongxue가 영화의 수 있습니다.
이 보고서는 중국의 자동차 산업에 대한 조사 분석을 통해 자동차 음성 시스템의 탑재 상황, OEM 및 공급업체, 산업 체인, 개발 동향 등에 대한 정보를 제공합니다.
목차
제1장 자동차용 음성 산업의 개요
자동차용 음성 산업의 개요
자동차용 음성 시스템의 탑재 상황
제2장 OEM의 자동차 음성 시스템 이용
SAIC
BYD
Changan Auto
GAC
Geely Auto
Jiyue
NIO
XPeng Motors
Li Auto
Leapmotor
Xiaomi
Harmony Intelligent Mobility Alliance(HIMA)
BAIC
Volkswagen
기타 모델
제3장 자동차용 음성 공급자
iFlytek
AISpeech
Cerence
Unisound
SoundHound
Desay SV
Mobvoi
Pachira
Huawei
Baidu
Tencent
Alibaba
SenseTime
VoiceAI
제4장 자동차용 음성의 산업 체인
PATEO
Tinnove
Megatronix
Haitian Ruisheng
DataBaker
Magic Data
Chipintelli
WUQi Micro의 자동차 등급 AI 음성 칩 : WQ5301
제5장 자동차용 음성의 개발 동향
OEM 기반 모델 구성 파라미터 요약(1)
OEM의 기반 모델 구성 파라미터의 요약(2)
동향 1 :
동향 2 : 사내 음성 개발 증가로 음성 제어는 점차 차량 이외에도 확대되고 있습니다.
동향 3 :
동향 4 :
동향 5 :
동향 6 :
KTH
영문 목차
영문목차
Automotive voice research: high-level voice function installation rate significantly increases, automotive voice moves towards "cognitive interaction"
From January to November 2024, installations of automotive voice systems reached 16.76 million units, with an installation rate of 83.3%. Compared to the full year of 2023, installations increased by 5 percentage points. By energy type, EREV (Extended-Range Electric Vehicle) had the highest installation rate for automotive voice systems, reaching 100% from January to November 2024. Typical models under this energy type include the Li Auto L series, AITO M series, and Deepal S series.
In terms of voice function, installations and installation rate for continuous dialogue, see-and-speak, and wake-up-free functions greatly increased in 2024.
For the see-and-speak function, from January to November 2024, its installations reached 4.66 million units, with an installation rate of 23%, an increase of 18 percentage points compared to the full year of 2023. By energy type, EREV had the highest installation rate at 92.1%, while fuel vehicles had the lowest at only 7.1%. By price range, the "see-and-speak" function had the highest installation rate in the over 500,000 RMB range, with representative models such as Zeekr 009, Yangwang U8, and NIO ES8. This range also saw the largest increase in installation rate, up by 48 percentage points. This also indicates a significant improvement in the intelligence level of automotive voice systems in 2024.
2. The cockpit accesses more ecological resources, voice assistants gain deep service capabilities
In the era of foundation models, a voice assistant that "knows a lot and can serve" relies more on the access to diverse ecological applications. For example, when users issue vague commands such as "the car is almost out of power," "I'm hungry," or "what should I wear for the Chinese New Year," the voice assistant's response requires support from applications like maps, local life services, and online information.
In addition to common applications like AMAP, iQiyi, Tencent Video, NetEase Cloud Music, and QQ Music, Li Auto has implemented voice calls to Xiaohongshu (Little Red Book) platform content and launched a deeply customized voice skill for Meituan. For example, users can wake up Lixiang Tongxue to ask " Chinese New Year outfits recommended by Xiaohongshu," "find a Beijing travel guide on Xiaohongshu," or "help me find a Cantonese restaurant on Meituan with an average price of 200 RMB and a rating above 4.5."
3. Foundation model applications accelerate the development of automotive voice from "command interaction" to "cognitive interaction"
Different from the previous command-based interaction, automotive voice systems empowered by foundation models have better capabilities in spoken language understanding, logical reasoning, knowledge Q&A, painting creation, and perceiving the vehicle's surrounding environment. For example:
XPeng's XGPT-powered Xiao P assistant has capabilities in spoken language understanding, logical reasoning, knowledge encyclopedia, painting & story & fairy tale creation, and recognizing objects around the vehicle.
Li Auto's Mind GPT-powered Lixiang Tongxue has fuzzy search capabilities, such as asking Lixiang Tongxue "I forgot the name of a movie, there's a black pianist, do you know what it is?"; search by image description, where Lixiang Tongxue can read movie poster content and express it freely, allowing children who cannot read to choose movies by describing the poster.
Xiaoai Tongxue's application of foundation models also gives it the ability to understand and respond to vague commands. For example, it can recognize and respond to commands like "Where is my phone?", "Turn off the lights at home", "What mountain is that ahead?", and "What car is that ahead?".
Taking XPeng Motors as an example, XPeng Motors has developed its own XGPT (Lingxi) foundation model and integrated it into the voice system. Additionally, it has integrated the ZhiPu AI base foundation model and multimodal models, giving the voice assistant Xiao P stronger language understanding, image recognition, and generation capabilities, which can be linked with in-vehicle perception system and external environment.
4. AI foundation models become a must-have for OEMs to build intelligent automotive voice systems
By 2024, the number of brands equipping their intelligent cockpits with foundation models has significantly increased, with Chinese independent brands being the primary drivers of this trend. Some brands have already completed the development path from cooperative supply to joint R&D, and finally to independent research. For example, in January 2024, Geely applied Baidu's ERNIE Bot foundation model in its Galaxy L6. In the same month, Geely released its self-developed full-scenario AI foundation model-Geely Xingrui AI Foundation Model.
Based on the Xingrui AI Foundation Model architecture, Geely has also developed derivative models such as the Xingrui NLP Language Foundation Model and the Xingrui Multimodal Foundation Model. Among these, the Xingrui NLP Language Foundation Model is entirely self-developed by the Xingrui Intelligent Computing Center, with a total training data volume exceeding 3 trillion tokens. It includes an emotional module, enabling excellent logical reasoning and contextual memory capabilities, allowing for human-like emotional interactions.
In January 2025, Geely showcased its development path for an in-cabin intelligent assistant based on the Xingrui AI Foundation Model at CES 2025-moving from "Assisted Intelligence" to "Agent Intelligence" and finally to "Autonomous Intelligence." With the support of the foundation model, in-car assistant will evolve from "accurately responding to commands" to "understanding the environment and autonomously completing tasks," and ultimately to "possessing self-awareness and autonomous emotional capabilities."
Chinese independent brands such as BYD, SAIC, Dongfeng, GAC, Changan, Chery, and emerging OEMs like NIO, Li Auto, XPeng, AITO, and Xiaomi have also implemented AI foundation models in automotive voice systems. As automotive intelligence enters its second phase, AI foundation models are gradually becoming a necessary option for building intelligent voice interaction systems.
Table of Contents
Related Definitions
1 Overview of Automotive Voice Industry
1.1 1 Overview of Automotive Voice Industry
Overview of Automotive Voice
Development History of Automotive Voice
Hierarchical Classification of Automotive Voice Interaction (1)
Hierarchical Classification of Automotive Voice Interaction (2)
Installation Forms of Foundation Model of OEMs
Key Participants in Foundation Model Installation
Enhancement Effects of Foundation Models on Cockpit Interaction
Pain Points in Automotive Voice Interaction Empowered by Foundation Models (1)
Pain Points in Automotive Voice Interaction Empowered by Foundation Models (2)
Pain Points in Automotive Voice Interaction Empowered by Foundation Models (3)
Other Voice Interaction Technologies-Voiceprint Recognition
Comparison of Voiceprint Recognition Applications by OEMs
1.2 Installation Status of Automotive Voice Systems
Overall Installation Status of Automotive Voice Systems
Installation Status of Advanced Automotive Voice Functions: Continuous Dialogue (1)
Installation Status of Advanced Automotive Voice Functions: Continuous Dialogue (2)
Installation Status of Advanced Automotive Voice Functions: See-and-Speak (1)
Installation Status of Advanced Automotive Voice Functions: See-and-Speak (2)
Installation Status of Advanced Automotive Voice Functions: Wake-up-Free (1)
Installation Status of Advanced Automotive Voice Functions: Wake-up-Free (2)
2 OEM Applications of Automotive Voice Systems
Summary of OEM Automotive Voice Interaction Functions (1)
Summary of OEM Automotive Voice Interaction Functions (2)
Summary of Foundation Model Applications by OEMs
2.1 SAIC
Application of IM Foundation Model in Automotive Voice Systems
Benchmark Model for Voice: IM L6
Benchmark Model for Voice: Rising F7
Details of Voice OTA Updates
2.2 BYD
Enhancement of In-Car Interaction by Xuanji AI Foundation Model
Benchmark Model for Voice: Denza Z9 (1)
Benchmark Model for Voice: Denza Z9 (2)
Details of Voice OTA Updates
2.3 Changan Auto
Deepal DEEPAL OS 3.0 Cockpit Interaction Capabilities
Avatr Harmony Cockpit Interaction Capabilities
Enhancement of Cockpit Interaction by Changan Xinghai Foundation Model
Benchmark Model for Voice: Deepal G318
Benchmark Model for Voice: Avatr 07
Details of Voice OTA Updates
2.4 GAC
AI Foundation Model Platform
Application of AI Foundation Model Platform in Intelligent Cockpit
Benchmark Model for Voice: AION RT
2.5 Geely Auto
Global AI Technology System for Intelligent Vehicles (1)
Development Path of In-Car Intelligent Assistant Under the Global AI Technology System
Application of Geely Xingrui AI Foundation Model in In-Car Assistants
Architecture of Geely Xingrui AI Foundation Model
Cooperation of Geely's Foundation Models
Zeekr Kr Foundation Model (1)
Zeekr Kr Foundation Model (2)
Structural Design of Zeekr Foundation Model
Flyme Auto Voice Interaction Capabilities
Benchmark Model for Voice: Geely Galaxy E8
Benchmark Model for Voice: Zeekr 7X
Details of Voice OTA Updates
2.6 Jiyue
AI Foundation Model Cockpit (1)
AI Foundation Model Cockpit (2)
Voice OTA (1)
Voice OTA (2)
Voice OTA (3)
Benchmark Model for Voice: Jiyue 07
Details of Voice OTA Updates
2.7 NIO
ONVO Intelligent Cockpit Interaction Capabilities
AI Foundation Model
NOMI GTP Interaction Framework
NOMI GTP Deployment Hierarchy
Nomi Module Design
Nomi Multimodal Perception Processing Flow
Nomi Command Reception + Scenario Understanding and Response Flow
Nomi Module Design-Command Distribution
NOMI Emotional Interaction Capabilities (1)
NOMI Emotional Interaction Capabilities (2)
Nomi Multimodal Perception and Interaction Capabilities
Nomi GPT Callable Scenarios
NOMI Vehicle Control Capabilities
Benchmark Model for Voice: NIO ET5T
Details of Voice OTA Updates
2.8 XPeng Motors
AI Xiao P
AI Underlying Capabilities
Enhancement of Voice Assistant Xiao P by XGPT
Remote Voice Vehicle Control Capabilities
Benchmark Model for Voice: XPeng P7i
Details of Voice OTA Updates
2.9 Li Auto
Empowerment of Cockpit Interaction by Mind GPT (1)
Empowerment of Cockpit Interaction by Mind GPT (2)
Empowerment of Cockpit Interaction by Mind GPT (3)
Module Design of Lixiang Tongxue Empowered by Mind GPT
Benchmark Model for Voice: Li MEGA Ultra
Application Scenarios of Mind GTP in Lixiang Tongxue
Skills of Lixiang Tongxue (1)
Skills of Lixiang Tongxue (2)
Skills of Lixiang Tongxue (3)
Skills of Lixiang Tongxue (4)
Details of Voice OTA Updates
2.10 Leapmotor
Empowerment of Voice Assistant "Xiao Ling" by Alibaba Cloud's Tongyi Foundation Model
Benchmark Model for Voice: Leapmotor C16
2.11 Xiaomi
Foundation Model Installation
Empowerment of Xiaoai Tongxue by Foundation Model (1)
Empowerment of Xiaoai Tongxue by Foundation Model (2)
Architecture Design of Xiaoai Tongxue
Module Design of Xiaoai Tongxue
Core Technology of In-Car Xiaoai Tongxue
Benchmark Model for Voice: Xiaomi SU7
Introduction of Foundation Model in Xiaoai Tongxue OTA
Details of Voice OTA Updates
2.12 Harmony Intelligent Mobility Alliance (HIMA)
Voice Highlights of HIMA Models (1)
Voice Highlights of HIMA Models (2)
2.13 BAIC
AI Agent Architecture
Benchmark Model for Voice: ARCFOX aT5
2.14 Volkswagen
Cockpit Foundation Model
Cooperation with Baidu on Voice Model Installation
Benchmark Model for Voice: Volkswagen ID.3
2.15 Other Models
Benchmark Model for Voice: Voyah Zhuiguang
Benchmark Model for Voice: WEY Blue Mountain
3 Automotive Voice Suppliers
Summary and Comparison of Automotive Voice Supplier Solutions (1)
Summary and Comparison of Automotive Voice Supplier Solutions (2)
3.1 iFlytek
Intelligent Cockpit Business Performance
Overview of Basic Voice Capabilities
Foundation Model Product System
Development History of Spark Foundation Model
Upgrade Content of Spark Foundation Model 4.0
Core Capabilities of Spark Foundation Model
Deployment Solutions for Spark Foundation Model
Vehicle Assistant Based on Spark Foundation Model
Spark Voice Foundation Model
Compatibility of Intelligent Vehicle AI Algorithm Chips
Multimodal Fusion Capabilities of Spark Foundation Model
Empowerment of Cockpit OS by Spark AI Capabilities
Low-Compute Deployment Solutions for Voice Algorithms
Case of Low-Compute Deployment for Voice Algorithms
Core Capabilities of Interaction Foundation Model
Application of Spark Foundation Model in Intelligent Cockpit
Summary of Experience in Deploying In-Car Interaction Foundation Models
3.2 AISpeech
Key Voice and Language Technologies (1)
Key Voice and Language Technologies (2)
Key Voice and Language Technologies (3)
Automotive Voice Solutions
Automotive Voice Assistant
Development History of Foundation Model
Details of Foundation Model
Architecture Deployment Diagram of Foundation Model
Design Diagram of Voice Foundation Model
Voice Development Platform
Full-Chain + Foundation Model Layout of Voice
3.3 Cerence
Automotive Language Foundation Model Solutions
Integration of Voice Assistant with Foundation Models
Voice Assistant
External Voice Interaction
Core Voice Technologies (1)
Core Voice Technologies (2)
Core Voice Technologies (3)
In-Car Interaction System
3.4 Unisound
AI Solutions
In-Car Foundation Model Solutions
Details of Foundation Model (1)
Details of Foundation Model (2)
Details of Foundation Model (3)
Business Model of Automotive Voice Solutions
Basic Voice Technologies (1)
Basic Voice Technologies (2)
3.5 SoundHound
Core AI Technologies (1)
Core AI Technologies (2)
Core AI Technologies (3)
Major Clients
Automotive Voice Solutions
Voice AI Platform
3.6 Desay SV
Research History in Automotive Voice
Application Scenarios of Foundation Model in Voice
Cloud and Vehicle Deployment Solutions for Foundation Model Voice
Automotive Voice Capabilities
Overview of Voice Foundation Model Solutions
Solutions to Industry Pain Points in Voice (1)
Solutions to Industry Pain Points in Voice (2)
Solutions to Industry Pain Points in Voice (3)
Solutions to Industry Pain Points in Voice (4)
Future Plans for Foundation Model Voice
Technical Reserves in Automotive Voice
3.7 Mobvoi
AI Voice Core Technologies
Intelligent Automotive Voice Solutions (1)
Intelligent Automotive Voice Solutions (2)
Foundation Model Solutions: Sequence Monkey (1)
Foundation Model Solutions: Sequence Monkey (2)
3.8 Pachira
Core Voice Technologies (1)
Core Voice Technologies (2)
Core Voice Technologies (3)
Voice Foundation Model Solutions
Intelligent Cockpit Foundation Model (Hybrid Architecture + Open Integration)
Automotive Voice Solutions (1)
Automotive Voice Solutions (2)
Advantages of Voice Products (1)
Advantages of Voice Products (2)
Highlights of Intelligent Cockpit Human-Machine Interaction Products
3.9 Huawei
Empowerment of Voice Assistant Xiaoyi by Pangu Foundation Model (1)
Empowerment of Voice Assistant Xiaoyi by Pangu Foundation Model (2)
Empowerment of Voice Assistant Xiaoyi by Pangu Foundation Model (3)
Empowerment of Voice Assistant Xiaoyi by Pangu Foundation Model (4)
Pangu Foundation Model 3.0 (1)
Pangu Foundation Model 3.0 (2)
Xiaoyi Dialogue Flow
Iteration of Pangu Foundation Model
Upgrade of Xiaoyi Based on Pangu Foundation Model 5.0
Commercial Vehicle Application Cases of Pangu Automotive Foundation Model
Qianwu Engine
3.10 Baidu
Apollo Super Cockpit Capability Framework
Apollo Super Cockpit Capability Framework-Fusion Perception (1)
Apollo Super Cockpit Capability Framework-Fusion Perception (2)
Apollo Super Cockpit Capability Framework-Fusion Perception (3)
AI Foundation Model Cockpit-SIMO 2.0
Operating System Solutions Based on ERNIE Foundation Model (1)
Operating System Solutions Based on ERNIE Foundation Model (2)
DuerOS X's Automotive Voice Solutions
Development History of Xiaodu Automotive Voice (1)
Development History of Xiaodu Automotive Voice (2)
Development History of Xiaodu Automotive Voice (3)
Empowerment of Xiaodu Assistant by Foundation Model
Intelligent Cockpit Foundation Model 2.0
Basic Architecture of ERNIE Bot Foundation Model
3.11 Tencent
AI Interaction Driven by Foundation Model
Intelligent Cockpit Foundation Model Framework
Applications of Intelligent Cockpit Foundation Model (1)
Applications of Intelligent Cockpit Foundation Model (2)
