The building blocks for the highly autonomous car are hugely complex, cemented in infinite lines of code, huge collections of data and artificial intelligence (AI). Bringing this all together is no easy feat, but it is one that Zenuity hopes to do with the help of its parent companies.
Established in January 2017, Zenuity is an equal joint venture (JV) between safety and electronics supplier Autoliv and Geely-owned Volvo Cars. Headquartered in Gothenburg, Sweden, it has facilities in Munich and Detroit. Zenuity specialises in developing software for advanced driver assistance systems (ADAS) and autonomous driving. According to Chief Executive Dennis Nobelius, Zenuity’s primary goal is to create “inspiring technology for safe and intelligent mobility.”
Speaking to Megatrends, Nobelius describes how the company is going about achieving its goal, and discusses the industry trends that are already having an enormous impact on the vehicles of tomorrow.
What makes Zenuity stand out from other companies that are developing ADAS and autonomous driving software?
I think one of the major strengths of our company is that we have all the intellectual property and patents from Volvo Cars and Autoliv in the field of active safety, and we also have many of the engineers that developed those technologies working for us. A third of our workforce is from Autoliv, a third from Volvo Cars, and everyone else is new.
Does this background allow you to act and react more quickly to certain issues or industry trends?
I’d say that we are working at a different pace from many other companies, and using different tactics. Every six weeks, we come together and decide whether or not the project we have been focusing on has reached the point that we wanted it to, and whether or not to take it further. Sometimes we change direction entirely and start a new project, other times we continue and make small tweaks. We’re very flexible and fast in our ideas and innovations.
Are you limited in terms of your customer base due to being owned by Autoliv and Volvo Cars?
That’s another thing that sets us apart from competitors. Even though we have all the experience and patents from our parent companies, we are free to work with any OEM that we like. This gives us the opportunity to work with the best hardware, for example, or fastest innovators in the industry.
What are the major trends currently governing advancements in ADAS and autonomous driving?
I’d say the first big trend is working with synthetic data, which is artificially generated data. The second is the use of artificial intelligence (AI). We realise that we must be effective with the data that we generate and handle, and AI is really transforming the way we use that data and, therefore, how we operate software.
How rapidly is AI developing in the automotive industry?
It’s happening very fast. We used to talk about software 1.0 – which is writing our own code from scratch. We’re now deep into software 2.0 – which means we are working with AI to produce and make sense of code. It’s the future, but it’s already happening now.
How could data be used differently to enhance vehicle design?
One of the big things we hope to see is the sharing of data between different OEMs. If they do this then the data becomes more valuable and progress with software development will speed up. But we need to be intelligent about what we share and with whom. So for example, slippery road alert is a really important feature in Nordic countries, and it makes sense to share this data to catch the dangerous spots. But it wouldn’t make sense to do this in Italy, for example.
The smart use of data and AI is vital in autonomous driving, and I think the higher the self-driving level, the more AI-based it will become. We are starting to see AI feature heavily in different software components to enable hands-off driving, taking in information from the hardware.
Are there any early examples of this?
Tesla’s latest range sensor is operated by neural networks, and I think we will soon see more examples related to driver monitoring, for example. These AI-based systems will be able to tell what the driver is looking at, and will be able to predict and offer information on what they would like to see, making the driving experience much more comfortable.
Will we ever see AI used in features such as braking and steering?
For safety-critical systems like steering it will come, but it will definitely take more time. We’re already doing object identification, using cameras and AI to tell the driver what they are seeing.
What are the challenges to consider when developing AI for automotive applications?
AI consumes considerable processing power in the car. A number of current evolutions will help, such as more powerful computers. Also, if you need to gather a number of petabytes of data with AI, it can become extremely costly and take a long time.
What are your expectations for the future of the self-driving vehicle market?
First of all, I think there will only be a few autonomous driving systems after 2023, because of the huge effort involved when making them. Only a few players will have the resources and capability to provide robust and safe systems, so there will be a streamlining of the industry. I don’t think a start-up with 500 employees will be able to go it alone, and so we will probably see much more in the way of partnerships too.
Do you think the rise of mobility services will have a big impact on the development of self-driving cars?
That’s a difficult question to answer, but I can envisage that there will be different types of autonomous cars used in cities compared to the ones used on the highway. Those operating in the city will replace taxis and will have a flexible layout, allowing users to stretch out and read, or access emails.
At the moment, however, we’re focused on self-driving vehicles for highway commuting. If you drive an hour to work each morning, 80% of that will be on boring roads, so the car could take over. It will still have a steering wheel, and the looks and driving feel will still matter. These factors will be less important for the autonomous cars in the cities.
Where do you see the automotive industry heading in the next ten years?
If you had asked me that 20 years ago, I would have been able to give you a precise answer. But today, there are so many disruptive trends that it makes it almost impossible to tell.
What we will probably see is more variety in transportation models. Today, the car looks like a car because of legal demands such as passive safety requirements. But with new technologies, we can totally remodel transportation, the Waymo self-driving bubble-like vehicles being a prime example. Forecasting and prediction will play a greater role in transport too. So, for example, if there is a big concert or event in town, the vehicles will be aware, and automatically reroute to ensure the traffic flow is more efficient. It’s a fantastic future where we are headed.
This article appeared in the Q1 2018 issue of Automotive Megatrends Magazine.