When people start listing innovations that will change the world, self-driving cars almost always get mentioned. As such, these vehicles are of key interest to car makers. In addition, this new technology is being viewed with mixed feelings. This is no surprise – conflicted responses have always surrounded transformative tech. This time, though, the potential changes are huge. Autonomous vehicles can disrupt current mobility concepts by:
  • Increasing productivity – instead of spending time driving, people will be able to engage in other, more important activities, which could include just watching a movie.
  • Reducing congestion – by properly coordinating vehicles, we can reduce stop-and-go waves and improve traffic flow.
  • Unlocking environmental and economic gain – people will see autonomous vehicles more in terms of mobility and less in terms of property, enabling a higher level of sharing and reducing individual cost as well as the total number of vehicles on the road.
  • Improving road safety – removing the human factor will reduce incidents resulting from distractions, intoxication, and the violation of speed limits.
Autonomous vehicles are coming to change society. Let us see how they work.

What is autonomous driving anyway?


Autonomous driving levels

“Autonomous” is a word that needs further clarification. When people talk about autonomous cars, they can be referring to functions ranging from lane correction to full self-driving capabilities. Thus, for clarity, the automotive industry talks about different levels of autonomy.


The classification currently used is SAE J3016 . It is based on the following criteria:
  • Does the software control both steering and acceleration/braking or only one of the two?
  • Is the software able to monitor the environment around the car to identify problems?
  • Is the software able to counteract the problems identified?
  • Does the software work only in specific scenarios or in all situations?
According to the answers, we get the following levels (green = vehicle software, yellow = human):

At the time of writing, there is no car on the market that meets level 3 requirements, regardless of what some marketing campaigns might say.
SW architecture, data processing, sensors

Empowering all levels above 0 are many lines of code and sensors that provide data. The higher the level, the more complex the architecture.


But, sticking to basics, here’s a general overview:

Dozens of sensors built into the car intake gigabytes of raw data including pictures, video, LIDAR point clouds, technical data from electrical and fuel-combustion engine control units, stabilization systems, safety equipment, and more. This data is processed, marked, and fused to provide all required information where and when it’s needed. This data can then further enable decisions that allow vehicles to perform their main function – driving safety – even more thoroughly.


Catching and processing this data is no easy feat, though. In fact, it involves the most cutting-edge tech such as image and video processing, mapping, localization, big data, machine learning, and more. Cars are now far from their mechanical origins. In fact, to some degree, they are now computers on wheels.

The role of software in automotive


The major recent revolutions in the automotive industry – autonomous driving, connectivity, electrification, and shared mobility, or ACES for short – have all been based on software.  Cars are now equipped with a network of electronic control units (ECUs) that are responsible for functions beyond those that are obviously code-based such as navigation systems or advanced driving assistance systems.


As we continue this journey of technological development, the software in the car will keep getting more and more complex. In fact, cars now have to handle gigabytes of information every minute reliably, safely, and in real-time. However, not only does automotive software lead to increased car complexity. It is now becoming a differentiating factor between different brands of cars.

Luxoft and its role in the autonomous driving landscape


Luxoft fuels this differentiation and enables automotive companies to excel. For several years now, Luxoft has been working on autonomous vehicles. We have developed know-how on the topic, and we have supported our employees as they have become experts in the area. Our development expertise was recently recognized by a premium German OEM that engaged us as a major collaborator for the development of the autonomous driving functions for an upcoming series of cars.


Our automotive software expertise is driven by our top-notch employees, so Luxoft is always looking for new talents. If you are interested, please check out our vacancies here.


Authors Thomas Labella Ph.D. , Aleksandr Golets, Evgeniy Uvarov
Thomas Labella
Thomas Labella is a manager with a history of working in the automotive industry. He has been at Luxoft for more than 3 years, and he has deep experience in autonomous vehicles, analysis, software design, and business case modeling. He is a strong program and project management professional with a Ph.D. in robotics and AI from Université libre de Bruxelles. He currently manages a department made up of more than 200 people.