Glen DeVos, Chief Technology Officer, Delphi:
On being a data broker
Are we more comfortable if the financial community starts to call us a ‘tech company’ and not an ‘automotive supplier’? Yes, in part. At the end of the day we still serve our OEM customers. But more value is being created outside those traditional portfolio elements. We want people to understand that we’re in that space as well. We’re not just a component supplier; we’re a data broker. That’s why we’ve made the investments in the past year, of ControlTec ([data analytics], Movimento ([software for OTA], Ottomatika [automated driving software], Ottonomo [data platform], as well as acquired stakes in Innoviz, Leddertech and Quanergy [lidar technology] and partnered with Intel and Mobileye.
One thing we like about Intel is they provide the ‘silicon’—the hardware-compute platform. And the software-enabled vehicle is all about that platform. They’re also in connectivity, data centers and the question we ask ourselves is, where should we be playing in that space? We’re not going to be a data-center operator; we need to add value. There are clearly opportunities for us to add value in the connected-vehicle space and in the cloud.
The traditional automotive ‘tiered’ supply structure is breaking apart a bit. It was very hardware-manufacturing-centric, stratified in many value-add steps beginning with raw materials. Much of the industry is still set up that way—we buy, source and procure in that model. But software doesn’t follow that model.
Guillame Devauchelle, VP Innovation & Scientific Development, Valeo:
On the changing role of R&D
We see implications of autonomous driving in every component we make, including wipers. Autonomy has made us focus on being more agile, running faster and ‘pushing’ the market. We believe newcomers such as Google and Apple are helping us to move faster. The automotive industry is the next ‘frontier’ of the Internet of Things because autos are now connected and autonomous. Our aim is to make the autonomous systems friendly and free of stress—you see the value and feel the benefits.
The role of R&D in the auto industry is changing. Some years ago it was all about engineering science, physics, mechanics and computing. Now it’s about adding value. For autonomous driving the car will not be autonomous 100% of the time. We need to understand when the driver is handed back control. So we need to understand biometrics, to monitor the driver.
Deep learning is a major component of our sensor fusion, for identifying a pedestrian crossing in front of me, for example. We have the processing power to make it happen. But to be safe we need triple redundancy in our autonomous system. Infrared technology is great but I am not sure if every car will have to have IR; it depends on cost. It has many wavelengths and each bandwidth has a specific ‘interest.’ We are developing it, but the cost is challenging.
Kathy Winter, VP Driving Systems Group, Intel:
On vehicle integration and autonomy
The Tier 1 role in vehicle integration is still important as we move into vehicle autonomy. What’s evolving is the high degree of software that’s in the vehicles. What used to be a very hardware-driven exercise is taking a lot of software expertise. So, an OEM’s choice of partners in this process is critical—who can you collaborate with to give a deep understanding, and optimization of, those software offerings, in addition to the traditional hardware? That key role may evolve over time but the basic job of integrating all the pieces doesn’t change. Tier 1s are truly expert at that, but as OEMs bring in more and more technologies and different pieces of software there may be a growing role for Tier 2s.
Tim Shih, VP Design, Yanfeng Automotive:
On his latest autonomous concept vehicle cabin
Our research for the past two years has focused on two main categories: working and being productive or relaxation ,and how that is enabled in the future. The only radical part of the ‘conventional’ driving mode, before the steering wheel folds and stows away, is the ‘transparent’ A-pillar. In reality there is a camera mounted in the blind spot with curved OLEDs [organic light emitting diodes] on the inner surface focused to the driver’s eye point. If a pedestrian walks through that blind spot the driver can easily spot them.
We have relocated the start-stop button and the PRND gear selector to the header rail. Not only does this free up storage space between the front seats but, once in autonomous mode, the driver doesn’t need them again. And there are different seating configurations in autonomous driving. In ‘Family’ mode, the rear seats close up to each other and the front seats rotate slightly inwards for easier interaction between the occupants. For ‘Work’ mode, the rear seats fold away; rather than waste space the driver’s seat moves rearwards and the passenger slowly rotates around to face them. There are also several working surfaces for lap tops, documents etc.
This makes it challenging when it comes to safety, but we have sensors throughout the interior to monitor where the passengers are located for optimally triggering the airbags.
Dr. Stefan Sommer, CEO ZF:
On the irrelevant 'tipping point'
We earn most of our revenue from mechanical parts but we already have high share in ADAS technologies, camera parts. We’re seeing a smooth transition to this and the market will change. It’s camera and radar technologies where we’ll be earning profits. But you also need the mechanical components for autonomous driving. It’s not only about artificial intelligence.
You need by-wire steering, brakes and ‘intelligent’ brakes capable of executing what artificial intelligence has thought through. In the future we see for autonomous driving and shared mobility, we still think the majority of our profits are still coming from mechanical components. You still need chassis, brakes and drivetrains, whether electric or combustion engine. But this will be enriched by intelligent technologies to serve the best functionalities. In the end it will be a balance between the new and old technologies.
I don’t see a ‘tipping point’ where the electronics and sensors part of ZF overtakes the mechanical/driveline part of the company. It’s not relevant for us. If we are successful in selling ProAI and radar technology we are happy to do that and grow our market share in this direction. In the same way we care our mechanical share. It’s not a company target to be bigger in electronics or big data. We see it all belonging together in ‘See, Think, Act’—and we need all of that.
In 10 years I see software being ZF’s biggest need, in terms of quantitative engineering resources. It’s the fastest growing technology in this area. But on the mechanical basis we still need resources on the same level as today. In all of the mobility solutions, whether private or shared, we need the same quantity and quality of mechanical engineering in our company. And in the full-autonomy future, we’ll likely need more sophisticated mechanical solutions.
Dr. Oliver Maiwald , Senior VP Powertrain Technology & Innovation, Continental:
On doing your homework
For us in Powertrain, autonomous driving is a ‘use case.’ There are plenty of options to utilize ICEs on many levels of automated driving, but the future for us is electrification. A lidar sensor requires 3kW–that’s equal to 6 g of CO2. So the ADAS guys have to do their homework to reduce the power requirements. Powertrain has to ‘pay the bill’ for CO2 emissions! We are constantly looking at what impact the new systems’ power requirements have on the board and the E/E architecture. I think for most people, however, safety trumps emissions.
We are striking a balance. We have to push electrification because of the stringent CO2 targets. That includes 48V on C-segment and larger cars and U.S. light trucks beginning next year, up to plug-in hybrids and battery electric vehicles. Everybody is painting a nice picture of battery electric vehicles, and we are committed to them.
But there are challenges that exist with the power grid being weak in many places, and the end consumer is very reluctant. And when government incentives are dropped, as has happened in Norway, the EV volumes drop, too. The point is now, how big is electrification? The car industry will push some technologies, and we have some homework to do on the hurdles.
PULL QUOTE: In 10 years software will be our biggest need, in terms of engineering resources. –Dr. Stefan Sommer, ZF