While the automotive industry is already firmly on the path to launching software-defined vehicles (SDVs) in the not-too-distant future, there are still a number of challenges to overcome. This is particularly the case for the more established automotive manufacturers, which are grappling with legacy technology, infrastructure and thinking. Meanwhile, their newer counterparts in countries such as China, which are unburdened by these legacies, are setting the pace towards a software-defined future using best practices learnt from their more established rivals.
AI has a pivotal role to play in helping to move SDVs forward and will eventually speed up the process, but right now, the industry needs to invest time and resources into understanding the impact and implementing it.
Generative AI will act as a powerful catalyst for legacy car makers as they transition towards delivering SDVs to market. Among the main benefits is improved efficiency. AI is good at mimicking what we teach it so is well placed to help with the more tedious tasks such as creating and intelligently populating a template configuration file with parameter data. It also facilitates enhanced design and optimisation. By analysing data and selecting the right parameters, generative AI can provide insights and suggestions for optimised functionalities for SDVs, leading to better, more efficient vehicles with improved performance and safety features. For example, an AI-based model for a Battery Management System may increase the vehicle range and make it more resistant to negative weather impacts.
In this way, generative AI can potentially help to bridge the gap between legacy automakers and the software-driven future by streamlining development, optimising designs and facilitating faster testing cycles.
There are many and varied applications for AI inside the SDV, where it will eventually influence everything from car maintenance and driving assistance to providing intuitive personal assistance. This includes applications such as advanced driver assistance and using AI to power fully autonomous autopilot capabilities, allowing for steering, navigation, route planning and parking in almost all environments. It also comes into play in monitoring vehicle health to identify potential issues, improve safety and reduce the chance of hold-ups.
Personalised in-car experiences will draw on AI’s capabilities, such as when planning routes in line with the driver’s schedule and learning driver preferences such as locations, temperatures or even audio entertainment. AI will also enable the driver to dictate an email or message and send it, having learnt the appropriate tone from the user. It also serves to reinforce security, ranging from biometric owner recognition to detecting suspicious activities and defending against cyberattacks.
The challenge we need to overcome is where the AI capabilities will be housed, in the cloud or in the vehicle itself. While the cloud might seem to be the obvious choice due to computation requirements, the vehicle would need a constant, uninterrupted connection with the cloud at all times. At this stage, that cannot be guaranteed and there are few signs that it would be possible. This would suggest that the more viable option would be to incorporate the processing ability within the vehicle itself.
The possibilities for AI in SDVs continue to evolve at speed and, with the SDV market expected to be worth almost US$250bn by 2032, it’s a huge opportunity for the automotive industry. The trick to making it a reality will lie in providing the platform and the AI-based software without compromising on safety of vehicle design.
The opinions expressed here are those of the author and do not necessarily reflect the positions of Automotive World Ltd.
Przemek Krokosz is Senior Solution Architect at Mobica
The AutomotiveWorld.com Comment column is open to automotive industry decision makers and influencers. If you would like to contribute a Comment article, please contact editorial@automotiveworld.comitect at Mobica,
