required

aimotive car
aimotive success

Great!

We have received your email.

aimotive failure

Something went wrong!

Please try again later or write an email to info@aimotive.com with the subject FWD TO: Info

Modular technology meets automotive supply chain needs

What’s the problem? Creating production-ready, compute-intensive AI-enabled hardware platforms for automotive is becoming one of the most innovative, dynamic and creative areas for the semiconductor industry.


February 28, 2019

However, there’s a problem: the target OEMs and Tier1s for these platforms aren’t waiting for the semis to come up with the complete, perfect solutions. They have had their own Resource and Developing working on every aspect of hardware, software and algorithms for these platforms for years, as the automotive industry responds to unprecedented change across the entire industry. And while none of them has anything like a complete solution developed solely by themselves, each one has their own technologies they want to use to give them a unique edge as the new leaders fight to emerge.

With competition and Resource and Developing investment already so high, the challenge is therefore to find new ways to deliver technology to these system companies, without trying to compete with what they have spent many millions already to develop.

We don’t know what they don’t know

The challenge for designing future generation computation-intensive hardware platforms is that there are way too many variables. The semiconductor industry doesn’t like that. It means that either the chips are too big and expensive in order to cover all possible requirements, or they are too specialized - so the risk is high they won’t be quite right for the market as it matures. And both scenarios mean the volumes will be too small to justify taking to production.

This isn’t just a hardware problem – it is just as challenging for software IP. If you develop a comprehensive solution, you have the luxury of controlling many of the internal design interfaces and communications between functional blocks. But that limits integration with other IP. And OEMs and Tier1s want to be able to blend technology IP together to create a solution that maximises the usage of their own in-house IP, expertise and investment.

Too many unknowns, too little time

With HAVs (Highly Autonomous Vehicles), and even the more evolutionary ADAS (Advanced Driver Assistance Systems) , there are many more unknowns. What algorithms will we be using? Can those algorithms be certified safe under all conditions? What sensors will be used? Whatever we implement, will it still be competitive in 3-4 years time (the typical time from design to production for automotive)? Can I upgrade the platform or algorithms during the vehicle lifetime?

All of this is further compounded by the sheer number and complexity of players in the automotive e supply chain, each of which is seeking to include their own value-add to maximise margins. Many are highly specialized, with complex relationships built over many years. You can’t simply re-invent supply chains to suit your great new solution, no matter how compelling it may be technically.

Modularity is key

One approach, used by AImotive for example, is to break down each of its technology product solutions in to a collection of finer grain blocks, and find ways to enable customers to configure them uniquely. Customers can then look at all the building blocks that comprise the product being offered,  and discuss which of them they already have their own solutions for, whether developed in-house or through other supply chain relationships. The engagement discussion then rotates around how to combine the capabilities of the customer with the technology developer, and what new relationships need to be forged within the supply chain to ensure minimum disruption.

The effect of presenting new technology IP in this way can be striking, with the discussion now starting with collaboration and teamwork, rather than “make vs buy”. In this way, trust is established much earlier in the engagement cycle, resulting in closer relationships with customers.

Understanding the bigger picture

However it’s not simply a matter of showing what’s within your products: you also need to show how well your products integrate into the overall system. This has been a challenge for many semiconductor vendors, who have struggled to acquire or develop credible systems knowledge.

One approach is to develop prototype systems that do the job using your technologies. AImotive, for example, has a fleet of vehicles in 4 locations doing tests daily just the same as the OEMs it aims to provide technology products to – though it has no intention of becoming an OEM. This experience has enabled a far more collaborative approach by sharing the “battle scars” of building real working systems, and how each tackles the myriad challenges associated with creating a real HAV. Shared experiences lead to closer collaboration, as mutual trust is built when deciding how best to work together.

New models for changing industries

The automotive industry is undergoing unprecedented change. Who knows who the winners and losers will be in five or ten years’ time? By embracing a far more collaborative, modular and flexible approach to product offerings, together with maximizing shared systems knowledge and experiences, stronger partnerships will be formed that will survive the inevitable challenges of one of the most exciting technological decades ever