Designed by data: Applying machine learning to automotive product development

Sometimes the smallest detail can have a profound impact on product design. An obscure correlation hidden away among vast chunks of data can hold the key to revolutionising the next generation of a product. And yet spotting these correlations with traditional methods can be virtually impossible sometimes.

In recent years, engineers and researchers have begun to tap into the potential of machine learning algorithms for spotting these hidden trends. What was once an academic topic has emerged as a viable development tool that can dramatically accelerate the product development process. In some cases, machine learning can slash the requirement for physical testing by as much as 90 per cent, by enabling the engineering team to home in on the most promising areas of development.

Dr Richard Ahlfeld worked on this concept as part of his PhD at Imperial College London and NASA. It was this experience that led to him founding Monolith AI, the first ever software company to be accepted onto the Advanced Propulsion Centre’s Technology Developer Accelerator Programme (TDAP).

Ahlfeld and his colleagues have developed a software platform that provides a user-friendly interface for training machine learning algorithms and analysing their results. Its potential applications range from the design of shampoo bottles to space rockets, but some of the most exciting are to be found in the automotive industry. Here, complex data from different testing sources can be used to predict a new vehicle’s behaviour on the track before even having to build it.

The TDAP programme has given the company access to the APC’s network of automotive clients. This proved invaluable in identifying routes to market and opening the doors to potential clients that a software developer wouldn’t typically have had the opportunity of engaging with. It also provided a total of £104,500 of grant support, without the need to forfeit any equity or give away any intellectual property.

“We were the first software company to go through the programme, so I think it was a step into the unknown to a certain extent, but it’s certainly paid off,” comments Ahlfeld. “Everyone was really supportive and the programme has proved extremely helpful. I would absolutely recommend the TDAP programme to any companies looking into it.”

Since joining the programme, Monolith AI has grown at an exponential rate with contracts now signed for major OEMs including BMW, Honda and Siemens. What began as Ahlfeld working with two other people now has 16 members of staff, including six PhDs and five engineers. Last year, the company was cited by Financial Times as having one of the most highly qualified workforces of any start-up in London. As Ahlfeld notes, “It’s been quite a transformation.”

Other Case Studies

A rounded view: Developing a ground-breaking life-cycle analysis model to deliver sustainable tyres.

How do you go about improving a product when its performance is governed by a complex set of sometimes-conflicting parameters? It’s a dilemma that tyre engineers already face when it comes to the trade-offs found in

Read Study

Going the extra mile: How APC funding helped a consortium led by Penso to develop a new range of lightweight bodies for commercial vehicles

The fleet market is one of the most promising applications for lightweight composite materials. As Daniel Hurcombe, managing director of Penso Consulting, points out, adopting carbon fibre materials on a high-end sp

Read Study