However, the emergence of connected vehicles and the generation of data from various sources, such as customer usage data and fleet data, have significantly changed the industry landscape. In the past, data needed to be actively generated; now, the industry leverages the massive amounts of data generated from these sources to train artificial intelligence models in real-time environments, thereby improving the efficiency of the development process.
Applications of Artificial Intelligence in the Automotive Industry
The impact of artificial intelligence on the automotive industry is complex. It is crucial to differentiate between various AI applications, such as machine learning models to enhance physical models, large language models to improve software development efficiency, and AI tools to assist in understanding and processing documents. These applications play a key role in all aspects of automotive development, from improving user interfaces to enabling autonomous driving.
One of the main challenges of integrating artificial intelligence (AI) into automotive development is managing massive amounts of data. Traditional data validation and processing methods are no longer sufficient. Instead, advanced infrastructure and platforms are needed to effectively process and filter this data. Despite these challenges, the advantages of AI are evident, significantly reducing the time and costs associated with manual processes. For example:
Data from a single data source can be used by different teams in multiple ways (e.g., calibration and software development).
The testing process can be optimized.
Development activities can be optimized and continue to run.
Applications of Artificial Intelligence in Software Development and Testing
In the field of software development and testing, artificial intelligence tools have been key tools for improving efficiency and reliability for many years. For example, AI can automatically generate test cases based on requirements, thereby reducing the time required for manual processes and mitigating potential errors. After all, AI can operate 24/7, but like any data-driven technology, the quality of its results and analysis depends on the quality of the data used.
Furthermore, artificial intelligence can analyze videos and generate synthetic data from real-world scenarios that are too expensive to measure physically, thereby improving test coverage and enhancing the reliability of test results. It also has significant value in generating test cases from requirements. Previously, this was a manual process where engineers needed to read and interpret requirements and then manually create test cases. With AI, this entire process can be automated, significantly reducing the time required and lowering the risk of errors or misunderstandings.
However, there are no shortcuts in software development and testing. The integration of artificial intelligence into safety-critical areas (such as autonomous driving) requires rigorous verification and testing to ensure reliability and safety. Therefore, currently, artificial intelligence is merely a tool to shorten development time and enhance confidence in test results.
The Future of Artificial Intelligence in the Automotive Industry
Looking ahead, the potential applications of artificial intelligence in the automotive industry are extremely broad. With continuous improvements in computing power and reductions in cost, the role of AI in automotive research and development and testing will further expand. For example, due to the limited computing power within vehicles, companies are exploring extending computing power to the cloud.
However, the industry must address regulatory and liability issues, particularly those involving autonomous vehicles. If an error leads to an accident, who should bear the responsibility? The vehicle manufacturer (OEM), the software supplier, or someone else? This is a complex issue that the industry must gradually resolve as autonomous driving technology advances.
This raises the question: how fast are we moving? Nobody wants to fall behind, but moving too fast can lead to a loss of control over what's happening, which is unacceptable when safety is involved. Because ETAS's work involves vehicle safety, we focus on how to make vehicles safe and reliable, so we need to maintain control over the output of our AI tools and rely on their results. Looking ahead, we believe a balance will be struck between leveraging AI to improve efficiency and ensuring safety and reliability through rigorous verification processes.
ETAS's Artificial Intelligence Solutions
ETAS is at the forefront of integrating artificial intelligence (AI) into automotive development, and our commitment to innovation ensures we continue to play a key role in the ever-evolving field of automotive AI. Our approach includes leveraging AI to process and optimize data, generate test cases, and improve model reliability. For example, we apply AI tools at both ends of the development chain:
Left side/Coding: Generative AI and chatbot tools assist developers in solving common problems such as AUTOSAR, C code generation, ARXML files, and basic software configuration when using the ETAS toolchain. In this process, the AI tools learn common problems, providing direct answers to known issues or helping human developers more easily handle undocumented problems.
Right side/Verification: The AI calibration kit optimizes the calibration process by leveraging vast amounts of data and incorporating real-world information. Traditionally, this was a time-consuming and manually operated process. By introducing AI technology, efficiency is significantly improved, and knowledge transfer is ensured, preventing knowledge loss due to the departure of key personnel.
Another example is the ETAS Embedded AI Encoder, which enables the safe, rapid, and efficient deployment of AI functionality on electronic control units (ECUs). By leveraging the tool's advanced capabilities, development time is significantly reduced, freeing up traditional resources previously used for generating C code. Furthermore, when integrated with ETAS's ASCMO, AI models applicable to multiple domains can be generated, providing a one-stop solution for model development and C code generation. Through collaborative work, these tools expand the embedded encoder ecosystem, enabling model generation even for non-AI experts.
Artificial intelligence (AI) and its tools are profoundly transforming the automotive industry, bringing unprecedented opportunities for efficiency improvements and innovation. As the industry continues to evolve, the deep integration of AI will play a crucial role in shaping the future of automotive research and development and testing. ETAS, by actively embracing AI technologies, fully demonstrates the transformative potential of these technologies in creating safer and more efficient vehicles.
