What will the application of AI large models bring to automotive parts?

The AI large model is becoming a new 'eye-catching package' recently.

Notably, in this new trend, there are various forms of AI large models being adopted by both domestic and foreign car companies such as Xingtu, Zhijie, and Mercedes-Benz, as well as many automotive parts, internet, and technology companies conducting in-depth development. From the perspective of automobiles, AI large models not only enter vehicles as new incremental components of smart cars but also empower vehicles by supporting changes in components.

Empowering component design

The main characteristics of AI large models include 'intelligence' and 'learning.' Although they are still in the early stages of development and are not as 'smart' as some online opinions claim, they are indeed an advanced productive force that is rapidly developing.

Recently, car companies like GAC and BAIC, which have launched AI large model platforms, have chosen to make AI large models one of the important components of smart cars. Therefore, in automotive component design, the assistance of large models is reflected in both hardware and software. In hardware design, designers used to have to read a lot of materials and search for numerous parameters of similar or related products in development tools, then develop based on new application goals, which is a time-consuming process. However, by applying the 'learning' capability of large models, designers can analyze, filter, and solve problems more 'intelligently' when establishing digital simulation models. They can also provide more optimized options for common design issues such as mechanical strength, material optimization, size selection, and simulation testing, which not only saves time in component design but also improves indicators such as material, performance, size, and lifespan, enhancing material utilization and even optimizing manufacturing processes and compressing the verification cycle of new products.

In terms of software, a typical car company's development team, relying on manually written code, might only write tens of millions of lines in a year. With the help of AI large models, not only can efficiency be improved more quickly, but rapid automatic code generation can also be achieved. It is evident that as AI large models continue to mature and their applications accelerate, they will not only restructure the systems related to software code but also bring about a transformation in the entire component design system. For example, by analyzing a large amount of vehicle operation data and charging data, component companies can optimize battery management systems and charging strategies to improve battery performance and lifespan. Additionally, large models can help companies design more efficient electric drive systems and energy recovery systems, thereby improving the range and energy utilization efficiency of new energy vehicles.

"The operating system pre-installed with AI large models will achieve the integration of intelligent driving, smart cockpit, and the entire central computing and intelligent connectivity, with self-learning, self-evolution, and self-growth capabilities," said Zeng Wenxiang, a researcher at the Northern Big Data and Artificial Intelligence Research Institute. He stated that AI large models play an important role in automotive design and R&D, assisting designers in generating ideas, conducting design validation and simulation by processing massive amounts of data, thus accelerating the vehicle development cycle. At the same time, AI large models can perform vehicle dynamics simulation, crash testing, fuel efficiency optimization, and other tasks, helping engineers improve design development efficiency and quality. Although AI large models are still in the development stage, their progress is rapid, and like operating systems, their ecosystem construction remains an important component and support for their development, which should be given sufficient attention to accelerate further technological breakthroughs in ecosystem construction.

A good 'partner' in intelligence

One of the core component systems of smart cars is the smart cockpit, which is also a space where AI large models can showcase their capabilities. In the automatic driving of smart cockpits, the role of AI large models is mainly to achieve continuous evolution and improvement of vehicle intelligent driving through extensive data collection, ultimately reaching the intelligent state of 'car and person as one.' This means that the intelligent driving habits of each car are 'learned, summarized, and improved' based on the actual driving habits of the driver.

"AI large models are deep neural network models that utilize massive data for learning and training. Their characteristic is to learn and train using complex computational structures based on the collection of huge amounts of data," said Yu Rongjin, a researcher at the Shenzhen Institute of Advanced Technology, in an interview with China Automotive News. He noted that AI large models have strong 'learning' capabilities, allowing them to learn rich patterns and features, thus possessing powerful generalization abilities to make relatively accurate predictions on unknown data. These characteristics are practically helpful for the evolution of automotive intelligence and for solving complex tasks in automotive intelligence, such as computer vision in perception systems, natural language processing, and voice recognition. The emergence and evolution of AI large models further accelerate the process of automotive intelligence, enabling better understanding and processing of perception information and intelligent decision-making in complex scenarios during autonomous driving, thereby enhancing the safety of vehicles in intelligent states.

Moreover, due to the high 'intelligence' level of AI large models, they are gradually being tried for the development of new vehicle models. For instance, during the design and development phase of new energy vehicle models, the role of AI large models is to deeply explore consumer needs and preferences, helping car companies design personalized vehicle configurations and special features to meet the personalized needs of different consumers.

In fact, AI large models can also optimize production processes and supply chain management. For car companies to achieve precise energy-saving and efficient production, AI large models can analyze a large amount of production data and supply chain information, helping manufacturers optimize production line layouts, process flows, and material distribution, thereby improving production efficiency and reducing costs. By simulating and optimizing supply chain management, manufacturers can achieve timely supply of raw materials and reasonable allocation of production plans, reducing costs, minimizing waste, decreasing inventory backlog and production stagnation, and improving production line utilization, capacity, and efficiency.

Providing support for the entire chain

"In fact, intelligence has already penetrated the entire automotive industry chain, providing a foundation for the application of AI large models," Zeng Wenxiang stated. He added that besides component design, vehicle model development, and intelligent driving, AI large models can also provide personalized after-sales services and user experiences in automotive sales and after-market segments. For example, by analyzing vehicle operation data and user feedback, potential problems and needs encountered by users can be identified and resolved in a timely manner, providing customized maintenance and service plans. Additionally, by utilizing large models to analyze vehicle fault data and maintenance records, predictive maintenance and remote diagnostics can be achieved, allowing for timely identification and resolution of potential issues, thereby improving vehicle reliability and user satisfaction.

Indeed, AI large models also play a role in automotive sales and marketing. By analyzing a large amount of consumer data and market trends, they can help car companies and dealers predict demand, formulate marketing strategies, and design customized products and services to increase sales and customer satisfaction.

Furthermore, AI large models empower the automotive industry mainly relying on computing power. In this context, the provision of computing power is also a new topic. "In the automotive field, Alibaba Cloud will promote the implementation of AI in specific application scenarios in the industry based on the Tongyi large model," said Li Qiang, Vice President of Alibaba Cloud Intelligence and General Manager of the Automotive Energy Industry.

"With the application of AI large models across the entire automotive industry chain, profound changes will be driven in the automotive industry from factories to markets," Zeng Wenxiang stated. He noted that from a broader perspective, AI large models are accelerating the automotive industry towards digitalization and intelligence, and this situation also places higher demands on industry practitioners. Professionals who can master digitalization, artificial intelligence, and understand automobiles, components, and even marketing will become highly sought after.