Steven Jenkins is the VP of Technology Strategy with Magna Electronics where he leads advanced roadmap and product development including research and partnerships within the ADAS Domain. Prior to that Steven was also CTO at Veoneer.
1. How do you see L2 ADAS adoption in the US market going forward?
Currently, adoption is slower in the US compared to other parts of the world. However, it is expected to increase steadily due to ongoing developments and the impact of regulations and safety considerations.
2. If we define L2++ as NOA (hands or eyes on) (highway and/or city), how much extra cost does this add in 2025 to the base price of a vehicle? What is the projection for adoption of such systems over the next five years?
The additional cost for L2++ is not as significant as one might expect, as much of the system is reused between levels. Additional sensors are required to meet performance requirements, but the increase is not as substantial as for L3. It is anticipated that the global adoption rate of L2++ will more than double in the next five years and will accelerate further in the subsequent five years.
3. Will all L2++ implementations move towards central compute?
Moving towards central compute is a logical step, offering numerous advantages, including the benefits of low-level fusion between sensors.
4. Magna announced a new ECU with Nvidia at GTC (Drive AGX Thor). Can you tell us more about that?
- How do you scale that from L2+ to L4 hardware-wise?
- The hardware scales based on the sensors and the processing power of the scalable SoC platform. This allows for a significant portion of the software to be reusable, although additional safeguards and larger AI models are necessary.
- The hardware scales based on the sensors and the processing power of the scalable SoC platform. This allows for a significant portion of the software to be reusable, although additional safeguards and larger AI models are necessary.
- What about the Magna L2++ and L3 software offering for OEMs?
- We offer both our own software stack and collaborate with specific third parties on the Nvidia platform. The software stack is a significant part of the system and has become more complex over the past 5-10 years. Therefore, we provide both software modules and work with third parties, including the customer’s own software. This approach offers customers the benefit of choice. As a Tier 1 supplier, we add substantial value by ensuring the final delivery and robustness of the entire system.
5. We discussed next-generation AEB at our recent Detroit workshop, but can you give us a summary of other L2 feature additions by 2029 that will use the same sensor set and compute in the base car configuration?
Development will focus on integrating both exterior and interior sensing on the same compute unit, enabling a more collaborative L2 system with an enhanced personalized experience.
6. For existing OEM platforms, is it possible to meet the new AEB standard with just software upgrades, or are hardware changes required? In Magna’s opinion, what are the minimal hardware upgrades needed from today’s AEB system?
There is no single answer, as it is highly dependent on the customer and platform. ADAS work in conjunction with lighting and braking systems to deliver the required performance characteristics. The ambition level also plays a role, whether to meet the minimum standard test or to achieve higher system availability with increased robustness against real-world variations. At Magna, we collaborate with our customers to meet standards and then scale to higher availability with our sensor and software offerings.
7. For new vehicle platforms coming to market by 2029, what is the next optional sensor that OEMs should have the capability to add to meet the new AEB standard?
Thermal sensors could enhance the system’s performance, increasing robustness and availability.
8. Thermal cameras seem to be the best solution, at least performance-wise, for AEB. What other L2 features do these enable? Will thermal cameras be limited to safety-oriented OEMs that want to exceed regulatory requirements, or do you foresee wider adoption?
Using sensors that provide broader coverage, and different modalities can enhance both the performance and availability of the system. To achieve better comfort and safety functions, sensors should operate effectively beyond ideal scenarios. Therefore, sensors like thermal cameras, which are not yet widespread in the vehicle fleet, can add value in various scenarios. For example, thermal imaging technology is beneficial for detecting wild animals that may appear from the side of the road with less illumination and occlusions. There is no reason why these sensors cannot complement vehicle perception to enhance the robustness of the feature set.
