自動駕駛模擬的產業鏈 (中國企業):2022年
市場調查報告書
商品編碼
1187317

自動駕駛模擬的產業鏈 (中國企業):2022年

Autonomous Driving Simulation Industry Chain Report (Chinese Companies), 2022

出版日期: | 出版商: ResearchInChina | 英文 140 Pages | 商品交期: 最快1-2個工作天內

價格
簡介目錄

本報告提供自動駕駛模擬的市場調查,自動駕駛模擬概要,中國企業的展開,各公司的平台概要,功能、特徵等資料。

目錄

第1章 自動駕駛模擬概要

  • 自動駕駛模擬技術概要
  • 自動駕駛的R&D中模擬實驗的意義
  • 自動駕駛模擬技術類型
  • 自動駕駛模擬軟體的結構
  • 模擬軟體系統概要圖
  • 自動駕駛模擬國際標準化組織:ASAM
  • 中國的自動駕駛模擬實驗標準的現狀
  • 中國參加的自動駕駛試驗情境的國際標準制定的現狀
  • 中國的自動駕駛功能模擬標準的發展
  • 中國的多情境多引擎模擬試驗服務平台解決方案
  • 中國的OEM與自動駕駛模擬平台的合作
  • 中國的自動駕駛模擬平台的比較

第2章 自動駕駛模擬的平台和企業

  • PanoSim
  • 51World
  • Huawei
  • Baidu
  • Tencent
  • 阿里巴巴(DAMO Academy)
  • IAE
  • CAERI
  • Saimo Technology
  • CICV
  • SMVIC
  • CATARC

第3章 自動駕駛模擬趨勢

  • 趨勢1
  • 趨勢2
  • 趨勢3
  • 趨勢4
  • 趨勢5
  • 趨勢6
簡介目錄
Product Code: FZQ006

Simulation Research (Part II): digital twin, cloud computing, and data closed-loop improve simulation test efficiency.

Simulation tests can not only be conducted in extreme working conditions and more complex scenarios and make ADAS/ADS verification more effective, but also reproduce and generalize the real vehicle test data, allow for deeper analysis of the problems in real vehicle tests and make corresponding optimizations, speeding up function development and shortening test cycle. The higher efficiency of autonomous driving simulation tests comes with the adoption of such technologies as digital twin, cloud computing, and data closed-loop.

1. Digital twin technology will help to build more extreme test scenario combinations.

Scenario libraries are the basis of simulation tests, and digital twin technology is a powerful tool for building virtual scene libraries. To ensure the safety and reliability of vehicles, OEMs need to test almost unlimited scenarios. By referring to the real world, digital twin technology can be used to model a 3D elements library quickly and automatically, and build different roads, marking lines, weathers, surroundings and other scenarios to achieve more possible test scene combinations, thus enabling high-precision simulation of sensors, environments, vehicle dynamic models, etc. Especially in the software OTA regression testing, digital twin can also greatly improve the efficiency of simulation testing and verification.

At present, Chinese comprehensive simulation platforms like Baidu, Huawei, Tencent and Alibaba, as well as specialist simulation testing service providers such as IAE, have all used digital twin technology for scene construction.

Huawei Octopus Platform can convert the collected typical road sections into simulation scenes, and combine them with HD maps to realize digital twin of real scenes. It can not only restore more than 95% scenes, but also give great assistance to developers to quickly simulate surrounding vehicles and realize minute-level scene construction. The platform with built-in 200,000 simulation scenes can provide application tools such as simulation, scene library management, scene fragment and evaluation system, as well as high-concurrency instance handling capabilities.

Tencent's autonomous driving digital twin simulation test platform TAD Sim (upgraded to 2.0) uses real data and gaming technology as dual-engine drive, covers simulation models such as road scene, traffic flow, vehicle sensing and vehicle dynamics, and supports OpenX and OSI international simulation standards. It offers more than 1,000 scene types, and can also generate larger-scale, rich scenes through generalization.

Founded in 2018, IAE is committed to building the world's largest simulation test scene workshop (massive scene libraries) with high precision, high confidence, high coverage and high freshness, and providing simulation scene data and SaaS (Scenario-as-a-Service). Its "Shuimu Lingjing" Scene Workshop is built according to the related Chinese and foreign intelligent connected vehicle industry standards, real roads and traffic behavior characteristics. With artificial intelligence and digital twin as underlying technologies, and the cross-platform and big data drive as the principle, the platform can be used to develop and build a whole-process and automated tool chain covering scene data collection, processing, analysis and mass production, realize large-scale, high-quality production of simulation scenes, and build a core support system required for large-scale algorithm training, simulation testing and evaluation. At present, IAE has built more than 8,000 groups of actually available simulation scene libraries, covering city-level digital twin, autonomous driving, Chinese and foreign regulations and standards, CIDAS traffic accident recurrence, safety of the intended functionality, and V2X.

2. The simulation testing based on cloud high-concurrency operation will further improve iteration efficiency of ADAS/ADS functions.

For advanced function development and intended functionality development, the autonomous driving simulation test platform needs to offer real restoration test scenes, make good use of collected road data to produce simulation scenes, and be capable of large-scale parallel processing on the cloud, so as to answer the needs of autonomous driving for closed-loop testing of perception, decision and control full-stack algorithms. Currently, technology giants, automakers, solution providers, and simulation software companies are working to expedite the construction of virtual simulation cloud platforms.

Baidu Apollo Simulation Platform is a cloud service built on Baidu Cloud and Azure. It improves the operating efficiency of simulation platforms through the large-scale distributed and dynamic variable speed simulation. Based on the large-scale cloud computing capacity, Apollo has created a virtual operating capability of millions of kilometers per day, and has built a fast iterative closed loop, making it easy for developers to achieve "millions of kilometers per day", greatly improving the development efficiency.

Alibaba Cloud Autonomous Driving Simulation Platform supports flexible, high-concurrency simulation and provides traffic flow simulation that can generate simulation traffic flows that conform to the element features and control methods of Chinese roads. Combined with autonomous driving simulation software, the platform enables game simulation, completing construction and testing of special scenes such as rainy/snowy weather and poor lighting conditions at night within 30 seconds. The Alibaba Cloud Platform favored 20 times faster autonomous driving simulation for Inceptio in 2022.

IAE "Jellyfish" Massive Simulation SaaS Platform can be deployed on private cloud and public cloud in a modular and elastic manner, and supports hypervisor, Docker and other modes. Besides designing and building cloud simulation platforms for customers, the company also builds a 400-node massive simulation SaaS platform based on proprietary cloud, with the virtual simulation test capability of daily effective mileage of more than one million kilometers, providing customers with SaaS-based simulation test services.

3. Building a data closed loop for autonomous driving simulation testing has become a new topic in the industry.

In the trend for "data-driven intelligence", simulation testing has become a key link in the autonomous driving data closed loop. How to build a data rolling iteration model through a range of simulation tests such as software-in-the-loop, hardware-in-the-loop, and vehicle-in-the-loop, and how to enable data-driven algorithm upgrades through corner cases in simulation tests have become new topics in the industry.

In March 2022, Tencent and Automotive Data of China (ADC) signed a cooperation agreement, under which data closed-loop and simulation testing for mass production becomes one of the R&D priorities.

In September 2022, IAE struck a strategic cooperation agreement with the autonomous driving industry data public service platform VDBP under the China Association of Automobile Manufacturers (CAAM). Through the close partnership with the CAAM and the VDBP platform, IAE will expand as many simulation scene data sources as possible, solve the problems of insufficient original data and single sources, and serve more Chinese and foreign OEMs relying on the platform.

In November 2022, Baidu announced a data closed-loop compliance solution for autonomous driving. Through the proprietary cloud platform, data decryption and data desensitization are carried out for simulation training, which ensures data compliance and confidentiality while implementing simulation testing.

IAE's X-IN-LOOP simulation test technology system integrates the concepts of technology closed-loop and data closed-loop throughout the entire vehicle development and verification process, and provides complete technical solutions and services from software/hardware-in-the-loop, driver-in-the-loop, advanced vehicle-in-the-loop, and vehicle-environment-traffic-in-the-loop to digital twin scene libraries and massive cloud computing power simulations, enabling the temporal and spatial acceleration of autonomous driving R&D, testing and verification to power the commercialization of autonomous driving.

In addition, from simulation objects, it can be seen that the trend for autonomous vehicle and V2X integrated simulation is accelerating. In current simulation software, road signs, marking lines, and road facilities act as static environment elements. As vehicle-infrastructure cooperation and Internet of Vehicles technologies advance, infrastructures such as road perception and communication will participate in the interaction of driving behaviors between autonomous vehicles, and the simulation of vehicle behaviors will pose new technical requirements as urban intelligent infrastructures work.

Table of Contents

1 Overview of Autonomous Driving Simulation

  • 1.1 Overview of Autonomous Driving Simulation Technology
  • 1.2 Significance of Simulation Testing to Autonomous Driving R&D
  • 1.3 Types of Autonomous Driving Simulation Technology
  • 1.4 Composition of Autonomous Driving Simulation Software
  • 1.5 Overview Diagram of Simulation Software System
  • 1.6 International Organization for Standardization of Autonomous Driving Simulation: ASAM
    • 1.6.1 Association for Standardization of Automation and Measuring Systems (ASAM)
    • 1.6.2 Chinese ASAM Standards: C-ASAM Working Group
    • 1.6.3 ASAM Standard Domains
  • 1.7 Status Quo of Autonomous Driving Simulation Test Standards in China
    • 1.7.1 Autonomous Driving Road Test Standards - National
    • 1.7.2 Autonomous Driving Road Test Standards - Provincial/Municipal
  • 1.8 Status Quo of the Formulation of International Standards for Autonomous Driving Test Scenarios in Which China Has Participated
  • 1.9 Progress in China's Autonomous Driving Function Simulation Standards
  • 1.10 Multi-scenario Multi-engine Simulation Test Service Platform Solutions in China
  • 1.11 Partnerships between OEMs and Autonomous Driving Simulation Platforms in China
  • 1.12 Comparison between Autonomous Driving Simulation Platforms in China
    • 1.12.1 Recent Developments in Autonomous Driving Simulation Platforms in China

2 Autonomous Driving Simulation Platforms and Companies

  • 2.1 PanoSim
    • 2.1.1 Profile
    • 2.1.2 Autonomous Driving Simulation Test Platform
    • 2.1.2 Composition and Functions of Products
    • 2.1.2 Products and Features
    • 2.1.3 xPilot Autonomous Driving Simulation Test Platform
    • 2.1.3 Components of xPilot
    • 2.1.3 Core Advantages of xPilot
    • 2.1.3 Application Scenarios of xPilot
    • 2.1.4 PanoDrive Driving Simulator
  • 2.2 51World
    • 2.2.1 Profile
    • 2.2.2 51Sim-One Simulation Platform
    • 2.2.2 51Sim-One Simulation Platform: Scenarios
    • 2.2.2 51Sim-One Simulation Platform: Perceptual Simulation
    • 2.2.3 51Sim-One Simulation Platform: Regulatory Control Simulation
    • 2.2.3 51Sim-One Simulation Platform: Cloud Simulation
    • 2.2.3 51Sim-One Simulation Platform: XIL
    • 2.2.4 Dataverse Data Platform
    • 2.2.5 51Sim-One 2.0
    • 2.2.6 Application Case: Shanghai Motor Vehicle Inspection Certification & Tech Innovation Center (SMVIC)
    • 2.2.7 Cooperation Events
  • 2.3 Huawei
    • 2.3.1 Profile
    • 2.3.2 MDC Platform: Cloud Training and Simulation Services
    • 2.3.3 "Octopus" Autonomous Driving Open Platform
    • 2.3.3 "Octopus" Autonomous Driving Open Platform: One-stop Autonomous Driving DevOps Capabilities
    • 2.3.3 "Octopus" Autonomous Driving Open Platform: Digital Twin and Virtual-Real Hybrid Simulation
    • 2.3.4 Octopus Autonomous Driving Cloud Service of Huawei Cloud
    • 2.3.4 Octopus Autonomous Driving Simulation Platform: Simulation Services
    • 2.3.4 Octopus Autonomous Driving Simulation Platform: Cloud + AI + Hardware and Software + Chip Combined Ecosystem
  • 2.4 Baidu
    • 2.4.1 Apollo Simulation Platform
    • 2.4.1 Apollo Simulation Platform: Scene Library
    • 2.4.1 Apollo Simulation Platform: Evaluation Criteria
    • 2.4.1 Apollo Simulation Platform: Version Iteration
    • 2.4.2 AADS
    • 2.4.3 Cooperation Events
  • 2.5 Tencent
    • 2.5.1 TAD Sim Autonomous Driving Simulation Platform
    • 2.5.2 TAD Sim Autonomous Driving Simulation Platform: Product Features and Core Capabilities
    • 2.5.3 TAD Sim Autonomous Driving Simulation Platform: Scene Restoration and Digital Twin
    • 2.5.4 TAD Sim Autonomous Driving Simulation Platform: Environment Simulation
    • 2.5.5 TAD Sim Autonomous Driving Simulation Platform: Sensor Simulation
    • 2.5.6 TAD Sim 2.0: Combination of Gaming Technology and Simulation
    • 2.5.6 TAD Sim 2.0: Architecture Upgrade
    • 2.5.7 Derivation of Autonomous Driving Simulation: City-level Simulation Platform
    • 2.5.8 Cooperation Events
  • 2.6 Alibaba (DAMO Academy)
    • 2.6.1 Autonomous Driving Layout of Alibaba
    • 2.6.2 Cloud-based Intelligent Simulation Test Platform
    • 2.6.3 Hybrid Simulation Test Platform
  • 2.7 IAE
    • 2.7.1 Profile
    • 2.7.2 X-IN-LOOP Simulation Test Technology System
    • 2.7.3 "Shuimu Lingjing" Scene Workshop (Massive Scene Libraries)
    • 2.7.4 Scene Data Production Closed-loop System
    • 2.7.4 Continuous Simulation Testing Based on the Scene Workshop
    • 2.7.5 "Jellyfish" Cloud Computing Power Massive Simulation SaaS Platform
    • 2.7.5 Customer Cases of "Jellyfish" Cloud Computing Power Massive Simulation SaaS Platform
    • 2.7.5 ADAS-AD-V2X Hardware-in-the-Loop (HIL) Products and Customer Cases
    • 2.7.6 Vehicle-in-the-loop Technology System
    • 2.7.7 Advanced Vehicle-in-the-loop (VaHIL) Simulation Laboratory and Customer Cases
    • 2.7.8 Generate Training Dataset Based on the Scene Workshop
  • 2.8 CAERI
    • 2.8.1 Profile
    • 2.8.2 Scene Library and Simulation System Integrated Solution
    • 2.8.3 Virtual Simulation Scene Library: i-Scenario
    • 2.8.4 Chinese Typical Driving Scene Library: V3.0
    • 2.8.5 Virtual Simulation Scene Generation/Conversion Software and Hardware-in-the-loop (HIL)
    • 2.8.6 Scene Data Processing Cloud Platform: i-STAR
    • 2.8.7 Scene Data Collection Equipment Platform: i-Collector
    • 2.8.8 Autonomous Driving Evaluation Software: i-Evaluator
    • 2.8.9 Cooperation Events
  • 2.9 Saimo Technology
    • 2.9.1 Profile
    • 2.9.2 Intelligent Connected Vehicle Simulation and Testing Platform: Sim Pro
    • 2.9.3 Autonomous Driving Function Cloud Platform
    • 2.9.4 Key Technologies and Experimental Environment & Facilities
    • 2.9.5 Cooperation
  • 2.10 CICV
    • 2.10.1 Profile
    • 2.10.2 Virtual Simulation Test Evaluation
    • 2.10.3 China Intelligent and Connected Vehicle Basic Data Service Platform
    • 2.10.4 Cooperation Events
  • 2.11 SMVIC
    • 2.11.1 Profile
    • 2.11.2 Autonomous Driving Simulation Test Laboratory
    • 2.11.2 Autonomous Driving Simulation Test Laboratory: Service Capabilities
    • 2.11.2 Autonomous Driving Simulation Test Laboratory: Scene Library
    • 2.11.2 Autonomous Driving Simulation Test Laboratory: Application of Trunk Autonomous Driving
    • 2.11.3 Sensor Simulation
    • 2.11.4 Automotive Intelligent Computing System Public Service Platform
  • 2.12 CATARC
    • 2.12.1 Profile
    • 2.12.2 Driving Scenario Simulation Platform
    • 2.12.3 Autonomous Driving Simulation Cloud Platform: AD Chauffeur 2.0
    • 2.12.3 AD Chauffeur 2.0: Core Features
    • 2.12.3 AD Chauffeur 2.0: Function Upgrades
    • 2.12.4 Scenario Generalization Tool: AD Scenario

3 Autonomous Driving Simulation Trends

  • 3.1 Trend 1
  • 3.2 Trend 2
  • 3.3 Trend 3
  • 3.4 Trend 4
  • 3.5 Trend 5
  • 3.6 Trend 6