Large AI Models and NOA Innovations Drive Transformations in the Automotive AI Algorithm Space

Dublin, Nov. 08, 2023 (GLOBE NEWSWIRE) -- The "Automotive AI Algorithm and Foundation Model Application Research Report, 2023" report has been added to's offering.

The landscape of the Advanced Driver Assistance Systems (ADAS) industry is undergoing rapid transformations, driven by the integration of large AI models and the emergence of Navigation on Autopilot (NOA) technologies. In 2022, ADAS companies were focused on introducing driving-parking integrated solutions, anticipating a boom in the market for 2023. However, the industry took an unexpected turn in 2023, with Original Equipment Manufacturers (OEMs) facing cost pressures and shifting their focus towards NOA solutions.

NOA Takes Center Stage:

In 2023, competition and innovation in the ADAS industry have pivoted towards highway NOA and urban NOA, thanks to companies like Huawei,, Baidu, and emerging automakers. Notably, a Southwest China-based OEM's highway NOA project, previously entrusted to medium-sized Tier 1 suppliers, faced challenges and was subsequently taken over by DJI and Huawei. Huawei, known for its work on high-end models, also ventured into low- and mid-end NOA solutions to compete with smaller ADAS Tier 1 suppliers.

Emerging automakers have also joined the race to introduce NOA technologies in various cities. In August 2023, Tesla showcased its AI-driven FSD V12, an end-to-end autonomous driving system, marking a significant milestone in autonomous driving. FSD V12 relies on large AI models, reducing the traditional codebase and achieving remarkable autonomous driving performance.

Impact on the ADAS Industry:

The industry now faces critical questions about its future direction and how to navigate the challenges presented by large AI models and NOA technologies. The "Automotive AI Algorithm and Foundation Model Application Research Report, 2023" delves into the history of ADAS algorithms and large AI models, offering insights into their future development trends in the automotive sector.

End-to-End Autonomous Driving vs. Modular Approach:

The report distinguishes between end-to-end autonomous driving and modular autonomous driving systems. The latter involves multiple layers for environmental perception, decision-making, and control. While modular systems benefit from division of labor, they can be complex and require significant manual design.

In contrast, end-to-end autonomous driving streamlines the process by directly feeding sensor data to a deep learning neural network, resulting in simplified rule sets and more efficient learning. This approach is expected to become mainstream as urban NOA gains prominence.

The Role of Large AI Models:

The transition to urban NOA presents new challenges and corner cases for autonomous driving models. Large AI models are essential to enhance generalization capabilities, reduce hardware costs, and eliminate the need for high-definition maps. They excel in handling complex urban scenarios, including unpredictable road conditions and multiple traffic participants.

Meeting the Data and Computing Requirements:

To leverage large AI models effectively, substantial amounts of data and computing power are required. Training Transformer models, for example, demands extensive mileage data—often exceeding 100 million kilometers. Real data and simulation scenes play crucial roles in model development. Supercomputing centers equipped with thousands of GPUs have become essential infrastructure for training AI models.

Industry Evolution and Opportunities:

While established automakers like NIO, Xpeng, and Li Auto have made strides in adopting foundation models, many other OEMs face challenges in terms of data and computing investment. Collaboration with large AI model providers offers a practical solution for OEMs to harness the power of these models.

For small- and medium-sized ADAS Tier 1 suppliers, independently developing competitive NOA solutions can be daunting. Industry integration is becoming inevitable, prompting some suppliers to consider partnerships, acquisitions, or going public to secure their future.

The disruptive changes in the ADAS industry also create opportunities for AI chip companies capable of customizing high-compute chips tailored to the requirements of large AI models. Emerging AI chip companies can play a pivotal role in advancing autonomous driving technology.

A Bright Future for the Industry:

The integration of large AI models and NOA technologies is poised to elevate China's position in the global vehicle intelligence landscape. Local Tier 1 suppliers in China are expected to partner with foreign OEMs and Tier 1 giants, further establishing their presence on the global stage.

As Amnon Shashua, CEO of Mobileye, aptly put it, "If you cannot win in China, you cannot win globally."

The ADAS industry is at a crossroads, with challenges and opportunities that will shape its future. Stay tuned for further developments in this transformative sector.

Key Topics Covered:

1 Classification and Development History of Autonomous Driving Algorithms
1.1 Classification of Autonomous Driving Systems
1.2 End-to-end Autonomous Driving and Software 2.0
1.3 End-to-end Autonomous Driving Model Case: UniAD
1.4 Development History of Baidu AD Algorithm
1.4.1 Development History of Baidu AD Algorithm: Model 1.0
1.4.2 Development History of Baidu AD Algorithm: Perception 1.0
1.4.3 Development History of Baidu AD Algorithm: Perception 2.0
1.4.4 Development History of Baidu AD Algorithm: Large Perception Model
1.4.5 Development History of Baidu AD Algorithm: Foundation Model Application Cases
1.5 Development History of Tesla AD Algorithm
1.5.1 Development History of Tesla AD Algorithm: Entering the Phase of "More Stress on Perception and Less Stress on Maps"
1.5.2 Development History of Tesla AD Algorithm: Occupancy Network
1.5.3 Development History of Tesla AD Algorithm: FSD Beta V12
1.5.4 Tesla Dojo Supercomputer

2. Common Autonomous Driving AI Algorithms and Models
2.1 Neural Network Models
2.2 Conventional Autonomous Driving AI Algorithms (Small Models)
2.3 Transformer and BEV (Foundation Models)
2.4 Algorithm Application in Different Scenarios
2.5 AI Algorithm's Requirements for Chips

3. Overview of Large AI Models and Intelligent Computing Centers
3.1 Overview of Large AI Model
3.2 Application of Large AI Models in Vehicles
3.3 Intelligent Computing Center

4. Tesla's Algorithm and Foundation Model Application
4.1 Tesla's Algorithm Fuses CNN and Transformer
4.2 Transformer Converts 2D into 3D
4.3 Occupancy Network, Semantic Segmentation and Spatiotemporal Sequences
4.4 LaneGCN and Search Tree
4.5 Data Closed Loop and Data Engine

5. AI Algorithm and Foundation Model Providers
5.2 QCraft
5.3 Baidu
5.4 Inspur
5.5 SenseTime
5.6 Huawei
5.7 Unisound
5.9 AISpeech
5.10 Megvii Technology
5.11 Others

6. OEMs' Application of Foundation Models
6.1 Xpeng
6.1.1 Profile
6.1.2 Transformer Model
6.1.3 Data Processing
6.1.4 Fuyao Intelligent Computing Center
6.2 Li Auto
6.2.1 Foundation Model Layout
6.2.2 Application of Foundation Models in Autonomous Driving
6.2.3 NPN and TIN
6.2.4 Dynamic BEV
6.3 Geely
6.3.1 Profile
6.3.2 Xingrui Intelligent Computing Center
6.3.3 Leading Technologies of Xingrui Intelligent Computing Center
6.3.4 Capabilities of Xingrui Intelligent Computing Center
6.3.5 Geely-Baidu ERNIE Large Model
6.4 GM
6.4.1 Consider Launching A Vehicle Voice Assistant Based on ChatGPT
6.4.2 AI Cooperation with Google
6.5 BYD
6.5.1 Utilize BEV Perception and Other Foundation Models to Overtake at the Bend
6.5.2 Multi-sensor Multi-task Fusion Perception
6.5.3 Large Data-driven Model That Will Realize the Whole Process of Perception, Prediction, Decision and Planning
6.6 Other Automakers
6.6.1 Great Wall Motor's Large AI Model Application and Layout
6.6.2 GAC Launches A Large AI Model Platform
6.6.3 Chery EXEED STERRA ES Carries A Large Cognitive Model
6.6.4 SAIC-GM-Wuling
6.6.5 Changan Automobile
6.6.6 Mercedes-Benz Applies ChatGPT

For more information about this report visit

About is the world's leading source for international market research reports and market data. We provide you with the latest data on international and regional markets, key industries, the top companies, new products and the latest trends.


Contact Data