There is no more recognized luxury car brand than Lamborghini. Lamborghini means power, style, and speed, but does it mean technology? Is autonomous driving on the horizon for the seminal Italian super sports car?
I had the opportunity to sit down with Lamborghini’s Director of Research and Development, Maurizio Reggiani, who deals with the development of long-term strategies, body and chassis technologies, and electronics for Lamborghini. Since January 2011, the Research and Development department, with Reggani at its helm, has also taken over the Lamborghini Style Center. Reggiani was kind enough to tell me all about the future of AI at Lamborghini.
How does Lamborgini see the future of autonomous vehicles, and the AI world?
I think we must first split Lamborghini into two branches: one is the super sports car, and the other one is our SUV, which will be launched in December this year. Speaking about the super sports car, our customers buy a Lamborghini because they want to drive a Lamborghini. They will never give the steering wheel over and leave the Lamborghini to something that is software, not themselves.
What I expect is that in the super sports car field, we will apply all the AI that is necessary to help facilitate the driver -- a system like adaptive cruise control that can help the driver use the car downtown or in a traffic jam. But the most important place to put AI is in software that can help the driver to improve the skill and activity of driving our car.
It’s important to us that customers can, as much as possible, use the car in an experienced way, but also stay safe and have the car under control if they try to push the limits of the driving experience. The car must be able to steer the driver to achieve the best use of the car.
You can imagine this classic example: you’re driving our car, with big tires and a huge amount of power, and on the corner you push it in wet conditions -- you’ll overspin the car very easily. It’s important that the car be able to give the driver the right instructions in real time, which allows the driver to arrive very close to the limits, but starts to give advice or some correction in a spontaneous way. The driver can then understand what happened, and learn about this to be as safe as possible. It’s like having a ghost driver, but more intelligent. The AI would not only be an indication of what is the best line you can follow, or the best gear, but it can work with you and be on your side to improve your skill, and in every condition be ready to close the gap between your skill and what the car is able to do. This is what I can imagine for Lamborghini AI in a super sports car.
To summarize, would it be safe to say that you see the autonomous system enabling you to drive the car as fast as you want, but preventing you from crashing?
It’s not just about going fast, but driving without losing control of the car. Sometimes you might take a corner at 30 miles an hour, and it’s already extreme. This is a classic example of how to control the car: to know when you must be cornering, where the points of tangency are, and where you must begin to steer. The AI will really make the driver able to experience the car. With AI you will know the exact angle of cornering, the points of starting and finishing, and based off the angle that you approach the turn, the AI can tell you open up or close the angle a little, or to be less aggressive on the accelerator. It can tell you to start to brake when you are straight, not on cornering, and to release the acceleration during the cornering to go faster.
The problem is applying this strategy without being too invasive, and to allow the driver to experience it in a positive, learning way, not in a panicked way. It’s easy to say, and I think it will be difficult to realize, but I think you must have a vision. Then engineers must work to apply it with researchers, sensors and all that is available on the market.
You’re obviously looking at AI in Lamborghinis, but can we expect to see autonomous driving in Lamborghinis in the future?
If you look at the Urus, our SUV that we will present at the end of year, it`s clear that we are going into a completely different field. Urus is a daily-use car, not like a super sports car which is mainly used on the weekend or a special occasion. In the Urus we will apply more AI systems that are able to support you in driving.
However, in the Urus we will try to avoid truly autonomous driving, because if you buy a Lamborghini, it’s still a Lamborghini. But we tried to introduce intelligent adaptive cruise control, and some other features that we think can support daily use of the car, without it being really autonomous. Autonomous driving, like everyone is talking about now, must remain out of game for Lamborghini. It will be difficult to motivate someone to spend $200K for a car, and then not want to use it.
Do you not think that when you are driving the car there are times when you want to drive the car yourself, but there are times when you are tired at the end of the day where might be nice to have the possibility to turn it on and have it take you home?
I’ve thought about this quite a bit. There is another big issue related to autonomous driving, and it’s that all the features that you need to implement in an autonomous car would kill our one important dogma, one important mission: weight reduction. Power is easy to increase, but weight is much more difficult to decrease. If you think of all the systems, all the features that are necessary to implement in an autonomous care like the camera, infrared, and so on, you can envision you’re adding more and more kilos to the car. Plus the impact that these components have in terms of packaging, volume, and design -- where will you store it?
The face of a car with autonomous driving is not a face; it’s like you have eyes in every place. Lamborghinis normally have a front that is like an arrow. To have an arrow, you must have a design, and the first reason to buy a Lamborghini is design. More or less, 60% of the value of the car is related to design, and is the difference between us and all the other competitors. Lamborghinis are cool, aggressive, innovative. Autonomous systems are more or less in a similar position on most cars, and this would be a problem in a super sports car. You can’t just apply systems from regular cars to a super sports car.
Another big reason is the reasons our customers buy a Lamborghini. When our customers are buying our super sports car, it’s not because they want to go from point A to B. It’s because they want to have fun, enjoy a weekend, have a day trip, and enjoy using the car, not because they need to use the car commuting from San Francisco to Monterey. They want to have fun.
Given that you could enhance the driver’s ability or limit the possibility of the driver making so many mistakes, do you see the possibility that in the future AI could make Lamborghinis drive faster, as they are less likely to crash?
With the super sports car, we have arrived close to the limit of components -- tires, power and aerodynamics. If you add AI based on the best lap time of the driver you might be able to complete this lap time again and again with the help of AI, but the first time it happens, it will be because the driver was able to do something in extreme conditions. Sometimes, a very good driver might be able to do something against physical law, something that hasn’t been done before. Afterwards, once you learn this from a driver, you could replicate it inside the software, and the AI will be able to do it every time based on the algorithm. But the software would never be the first one to do what a driver can do in terms of correction, or personal skill, which is really something AI can never have. It’s the difference between human and artificial, every time.
I’ve read that in Germany they’ve proposed to make all cars electric. Given that a large part of the experience of driving your cars is the sound, how do feel about that and how will you work with the sound branding element?
In the DNA of the super sports car, design is the first factor, second is the dynamic behavior of the car, the handling, third is sound, and then performance.
But before the sound I think there is another huge problem with electrification, which is when this kind of platform might be applicable to a super sports car. There are two fundamental problems: one is the weight, and the second is packaging. Super sports cars must have the weight as close to the center of the car as possible in order to maintain handling, the design would change drastically, and both points at the moment don’t have solutions.
The third problem is how much it costs -- I don’t know how many cars that are using full electrification are profitable. For a small company like us, we would lose money. In terms of timing, electric cars wouldn’t be possible before 2025 or 2027 in a super sports car.
But when it will arrive, and sooner or later it will come, sound will be one of the most difficult problems to solve. We’ve tried to look at artificial sound, with good sound engineering like a V12 engine. But I don’t think it’s something that our customers would appreciate, and for one simple reason: an artificial sound is the same every time. A real Lamborghini sound has definition: the RPM, the acceleration through a curve, the transitions. After ten times of hearing an artificial sound, it will annoy you and you’ll say, “Turn it off.”
The question for me is, in what way can we give a driving experience that is related to the sound? And I don’t know what we could build that would mimic that in a great way. It must be something that is related to the functionality, speed and power of the car. If it’s not related, then it’s better to have no sound at all -- but then in that case it isn’t a super sports car anymore.
Premium brands will try to be as silent as possible, so we need to invent an emotional connection to the sound of an electric car. We’ve already started discussing a crazy idea: theoretically, we could be using the sounds of the wind. The speed of the wind, the flow of the wind, related to the real dynamic use of the car.
Because the sound is really important for the Lamborghini brand?
The sound is one of the five most important factors for the DNA of the car, so we will need to find a way to recreate this.
Is there any way that you think AI will change our relationships to the automotive world that you’ve been considering in a different way?
It’s clear that the car will change dramatically, with packaging and electrification. The more that cars move into electrification, the biggest risk is that all the cars will be similar to others. Because then it will be the battery that will make the difference. It will be about how much energy you can put into the battery, what is the difference between the top peak and the average output of power? All of the cars must have the same speed. And for us, it’s all about what difference we can make between us and the other.
I believe today Toyota announced that it was going to invest in and launch a flying transportation mobility device. When we are we going to see a flying Lamborghini?
I think never. It comes back to the DNA: we need to have a wide tire, perfect friction, the perfect way to discharge the power of friction on the road. I think in this moment there is so much work to do on the ground movement. You mention the example of Toyota: Toyota is massive production, and they’re thinking of a movement. For us, it is different; we must think of what people will enjoy. This is the difference between us.
Again, if you’re buying a super sports car, you don’t buy Lamborghini because you need it. You want to give yourself an award, you want to gratify yourself. You want to say I have this car in this shape and this color, and it can do 0-100 km/h in 2.9 seconds. Is it uncomfortable? Yes! You have another type of car in the garage for day-to-day use. We think outside of the box and do not follow the mainstream brands. In a fully autonomous car, you’re going to be sitting in the same car, but in a different shell. This is not what our customers want. We sell emotion!
