Another notable change is the related upgraded offerings in computing power, necessary to support the intense requirements of artificial intelligence.
What do you see as the key technical challenges and opportunities facing automotive engineers in 2025?
I’ve noticed that development cycles in Europe are much longer than those in China, due to the EU countries’ more conservative approach to engineering. The European automakers missed out on pioneering electric vehicles—will they also miss out on autonomous vehicles?
How have software and the widespread adoption of AI changed your approach to automotive engineering and product development?
The Imagry approach has always been to utilise neural networks when developing our computer vision technology, and then our motion planning technology, the two core components of our autonomous driving software solution. AI-based systems, rather than rule-based systems, are now mainstream for autonomous driving purposes.
Thanks to AI, the learning curve is shorter. The push for AI across many applications has certainly influenced the chip manufacturers to create more powerful processors to support AI functionality within the vehicle, also known as edge computing. Without those advancements, it wouldn’t be possible to develop real-time, self-driving, mission-critical systems based on AI.
From low-volume, high value right up to mainstream automotive, engineers are under pressure to reduce cost and development time, while maintaining or even improving quality. How do you approach this challenge?
First, AI-based systems save vehicle training time. Additionally, the Imagry approach is HD-mapless, and LiDAR-free. Both HD maps and LiDAR are very expensive. In addition to their high price tags, the former requires continuous updates which mandate broadband communication with the cloud, while the latter cannot be justified for use in economy models, and furthermore, requires additional computing resources, adding cost.