Is AI the Future of Automotive Design?

The rise of AI (artificial intelligence) in recent years has been absolutely game-changing around the globe. AI has massively taken hold around the world in a wide variety of ways. One way you should immediately recognize is the featured image of this article, seen below, which EVinfo.net created in Google Gemini. See more at BillPierce.net. I offer these for free use, and appreciate attribution if you use them. However, attribution is not required. Thanks to Google and AI, I can now add automotive design to my resume.

AI is a branch of computer science focused on developing machines capable of performing tasks that traditionally require human intelligence. The field has seen rapid advancements, with recent developments enabling systems to outperform humans in some areas within just a few years. The global AI market is experiencing significant growth and is expected to continue expanding in the coming years. In 2023, the market size is valued at $196.63 billion and is projected to reach $1,811.75 billion by 2030, reflecting a compound annual growth rate (CAGR) of 36.6%.

Governments are making substantial investments in AI research and development, while leading companies are acquiring AI-driven startups. Around 83% of businesses regard AI as a top priority, and nearly half of them use it to analyze big data. AI is being applied across various industries such as finance, healthcare, retail, automotive, and manufacturing. In particular, generative AI is gaining traction, with one-third of organizations already incorporating it into their operations and 40% planning to increase their overall AI investments. Additionally, AI marketing is on the rise, with the AI-enabled advertising and media sectors leading the market in 2023.

MIT Engineers Focused on EV Designs, Recognizing Our All-Electric Transportation Future

MIT engineers have developed over 8,000 electric vehicle (EV) designs that can be combined with artificial intelligence (AI) to accelerate the future process of car manufacturing. Known as “DrivAerNet++,” this open-source database includes 3D models of car designs based on popular vehicle types, providing detailed information on aerodynamics and other specifications.

While electric cars have been in existence for more than a century, their popularity has surged in recent years, most notably in China, which recently passed 50% EV adoption. Traditionally, designing electric vehicles is a lengthy process, requiring years of work, iterations, and revisions to finalize a design before creating a physical prototype.

However, proprietary nature of these designs often means that test results, including aerodynamics, are kept private, which slows progress in improving EV range and fuel efficiency. The new database aims to expedite this process. By combining this rich collection of design data with AI models, manufacturers could rapidly develop new, efficient electric vehicle designs. This integration of AI could streamline a traditionally slow and resource-intensive process, enabling faster design and prototyping.

The MIT team outlined their work in a paper uploaded to the preprint arXiv database in June 2023 and presented it at the NeurIPS conference in December. Mohamed Elrefaie, a mechanical engineering student at MIT, emphasized that this dataset could reduce research and development costs, speeding up advances in the EV sector, which would benefit both manufacturers and the environment by bringing more efficient vehicles to market faster.

A key factor in this accelerated design process is the use of AI tools. The dataset enables the training of generative AI models that can perform design tasks in seconds instead of hours. While previous AI models generated seemingly optimized designs, they were limited by insufficient training data. The new dataset provides more comprehensive data, allowing AI to generate novel designs or assess the aerodynamics of existing ones, all without needing a physical prototype.

The team’s efforts produced 39 terabytes of data and consumed 3 million CPU hours using MIT’s SuperCloud, a powerful cluster of computers for scientific research. The engineers ran algorithms that adjusted 26 parameters—such as vehicle length, underbody features, wheel shapes, and windshield slope—on each baseline model. These designs were tested for originality and converted into various formats, including meshes and point clouds. Finally, advanced fluid dynamics simulations were conducted to assess airflow around the designs, providing valuable insights into efficiency and range.

(Image credit: Mohamed Elrefaie)

Most Americans don’t realize yet that the world’s transportation future is all-electric. The MIT team wisely realized that by making EV designs with their AI. Anyone looking into China immediately realizes it, as China recently passed 50% EV adoption. Sadly, most Americans are stuck in our dirty, gas-powered past, and threaten our automotive future by opposing the EV tax credit. See why EVinfo.net encourages saving the credit along with many jobs and our economy.