FISITA World Mobility Summit speaker Richard Ahlfeld, CEO at Monolith, shares his thoughts on the theme of the event, “To EV, or not to EV?”
Dr Richard Ahlfeld is CEO and Founder of Monolith, a no-code AI platform used by the world’s leading automotive, aerospace and industrial engineering.
With a PhD in Aerospace Engineering and Data Science from Imperial College, Richard has been recognised across academia and industry, including being named to MIT Technology Review’s Top 10 Innovators under 35, invited as a Research Engineer to NASA where he worked on the Mars rocket, and recognized as an Automotive News Europe ‘Rising Star’.
The Monolith platform leverages clients’ test data to build highly accurate, self-learning AI models. Used by engineers to reduce testing, increase learning, and maximise product quality, the platform has demonstrated the ability to decrease battery testing time by up to 73% using the latest robust active learning algorithm.
In the build-up to the 2024 FISITA World Mobility Summit, we asked Richard to share some thoughts on the theme of the event, “To EV, or not to EV?”.
The adoption of artificial intelligence is considered imperative in all sectors of industry. What have you identified as the key applications of AI in the automotive industry, and how has it been most successfully implemented?
AI is ideally placed to help engineers reduce time to market for new products, improve product quality, and reduce costs in the development process. At Monolith, we’re consistently seeing the most traction in material design and testing, from end to end.
Engineers find AI incredibly adept at helping them find better chemical formulations for higher-performing batteries. There are countless ways in which batteries can be configured, making it almost impossible to identify the optimum chemistry through physical testing processes alone. AI negotiates these intractable problems and identifies the best solution.
AI can also optimise test plans for more efficient programmes, inspecting huge datasets containing test results that are being collected every day. This avoids engineers needing to collect data manually, and catches errors early.
Furthermore, there is a significant role for such software in optimising the process for powertrain system calibration, which has historically required considerable data collection from field testing and dynamometer testing. Analysing field data collected from fleet telematics goes a long way to reinforcing this, but also to understanding driving behaviours and system performance over extended periods.
What are the key challenges for the automotive industry in adopting AI tools, and how could these challenges be overcome?
The automotive industry holds numerous long-standing testing and validation processes as the norm, with scepticism in the industry when it comes to changing these tried-and-tested methodologies. Transitioning from the usual method of physically testing everything, to accepting predictions as part of the design process with less testing, is viewed as dangerous in the eyes of many engineers, with fundamental trust issues over model accuracy.
Innovation needs to be encouraged, and we can overcome this by slowing the transition, enabling engineers to feel comfortable, and to become self-sufficient with machine learning tools and methodology. It’s imperative that machine learning expands beyond data science teams and is applied through engineers themselves.
The rate of battery technology development is at once remarkable in how far it has come, and in how much more it can be improved. What role can AI play in making notable advances in battery technology R&D?
There is huge scope for AI to make incredible efficiencies in battery material design. Significant advancements in material research and cathode design are already being made, and AI can narrow down material formulations to the highest-performing options, streamlining millions of combinations into hundreds of viable high-performing choices. AI can guide engineers to designs that may be better, leading to faster development of more productive prototypes. This means they can innovate more within the same development cycles, which facilitates delivery of better batteries, faster.
Aside from its deployment in R&D, how close are we to seeing AI used onboard the vehicle, in the use-phase? For example, in battery management systems for thermal management or for driving pattern analysis and power demand prediction?
AI already plays a large part in modern car operation. People are increasingly seeing AI in action with self-driving applications as part of everyday driving. Advanced driver assistance systems are fundamentally reliant on AI, even with simple features like lane-drift detection and object detection.
Battery state of health can also be more accurately estimated using AI, particularly thanks to higher-quality drive cycle analysis, and smarter charging patterns can be mapped, helping to preserve battery performance in the longer term.
You’re speaking at the FISITA World Mobility Summit in November 2024. The event is titled, ‘To EV, or not to EV?’ With that in mind, what will be your key message to delegates?
In the current market, I feel there is a clear need to build better EV and ICE technologies while we can, as hybrid-powered options will become hugely popular. Given increasingly stringent decarbonisation requirements in markets across the world, however, we also want to develop better battery EVs. While the technology is still relatively new compared to ICE vehicles, the industry at large is investigating sodium- and solid-state batteries to unlock more performance, durability, and cost-benefit. There’s much to be researched, discovered, and implemented in the race to make EVs more efficient, and AI is ideally placed to help engineers achieve that mission.