An autonomous vehicle isn’t just a massive, oddly shaped computer on wheels. It’s more of a data center made up of networked computers performing or responsible for different tasks. Naturally, an autonomous car operating system (OS) differs on these computers. Let’s find out which automotive OSs are critical to develop, deploy, and manage autonomous vehicle software and hardware.
In this article, you’ll learn about:
- What operating system do autonomous cars use?
- What are the top autonomous car operating systems?
- The current state of the automotive operating systems market
- The working principles of operating systems in autonomous vehicles
- The car operating systems favored in the automotive world
- The approach to developing automotive operating systems for autonomous vehicles
- 1 Complexity is the key attribute of automotive OS platforms in autonomous cars
- 2 Operating systems for autonomous vehicles that are technological monsters with multiple brains
- 3 Two approaches to automotive operating systems that automakers and OEMs may pursue
- 4 The bottom line
Complexity is the key attribute of automotive OS platforms in autonomous cars
Building an autonomous vehicle is a challenge. Building an autonomous car operating system is even more challenging.
Making cars smarter and safer is much harder than constructing any of the vehicles driving on our streets today. The same goes for automotive operating systems for autonomous vehicles. Choosing an automotive OS is a big deal, since it needs to control the car’s core capabilities while keeping passengers and the driving environment safe.
Building and running operating systems in autonomous vehicles requires not only knowledge but also enormous amounts of custom software development and computing power. Every single self-driving car uses hundreds of millions of lines of code. Check out this visualization comparing how many lines of code different services and apps need.
Lines of code required by vehicle OS, different services, apps, and other technologies
Challenges accompany self-driving technology through its entire development route. That’s why autonomous cars haven’t arrived en masse yet. Operating systems in autonomous vehicles should help automakers cope with the complexity and range of instantaneous tasks as well as with maintenance. What are the current OS platforms in autonomous cars?
Operating systems for autonomous vehicles that are technological monsters with multiple brains
This isn’t surprising, since they must substitute for the abilities of a human brain behind the wheel. This is a case when quantity is as important as quality. Electronic control units (ECUs) act as the brains for autonomous vehicles (AVs). ECUs can be perceived as minicomputers. They vary in size, purpose, and the operating system they run.
Electronic control units inside a vehicl
An autonomous vehicle is a diverse and complex environment that allows several operating systems to coexist. More than that, they can even run on top of one another on the same hardware. Every component, from the engine to the windshield wipers, and every car control, from steering to braking, is under the guidance of an ECU.
We can split all ECUs in autonomous vehicles roughly into two categories:
- ECUs responsible for ADAS and controlling car operations like steering, switching gears, braking, fuel consumption, etc. These ECUs can be managed by various real-time operating systems (RTOS).
- ECUs responsible for A/V infotainment systems and running different applications. These ECUs can be operated by operating systems more similar to those that run on a PC. These can be not only RTOS but also general-purpose operating systems (GPOS).
Now we’ll discuss the operating systems for autonomous vehicles from both of these categories that are favored by automotive developers.
QNX Neutrino is a recognized leader among OS platforms in autonomous cars. QNX Neutrino is a real-time operating system for cars used by automakers to run all processes and actions within the time frame needed for successful and safe execution. Along with its automotive OS, QNX Software Systems, owned by BlackBerry, builds advanced automotive development platforms, runs autonomous vehicle innovation centers, and offers dedicated teams (as used by Ford) to expand the QNX Neutrino OS and provide other services.
BlackBerry’s QNX claims that they’re building a connected car operating system along with secure, scalable, and trusted software solutions that enable the future of connected and autonomous cars. To prove it, the company invests heavily in new automated control systems for driverless and semi-driverless vehicles along with new safety and security systems and V2V and V2I communication technologies.
Automotive brands that use it: QNX Neutrino is already partnering with about 40 automakers including Ford, Acura, VW, BMW, and Audi. You can read more about QNX initiatives for connected and autonomous cars here.
VxWorks is a real-time vehicle OS that conforms with the ISO 26262 safety standard. Its creator, Wind River, is a global IoT software provider for safety-critical domains including automotive. This vehicle operating system helps automakers and OEMs deploy safe, secure, and reliable autonomous platforms.
Automotive brands that use it: Toshiba, Bosch, BMW, Ford, VW, and others. You can read more about Wind River initiatives for connected and autonomous cars here.
Green Hills INTEGRITY®
This real-time operating system for cars is part of the Green Hills Platform for ADAS, a complete solution for building the software components of secure, high-performance, and scalable ADAS systems. Green Hills claims that production-grade electronic control units powered with the INTEGRITY vehicle operating system will streamline the development and testing of life-critical autonomous vehicle applications.
Automotive brands that use it: Green Hills works mostly through partnerships with OEMs and Tier 1/Tier 2 providers. You can read more about Green Hills initiatives for connected and autonomous cars here.
NVIDIA DRIVE™ OS
Although NVIDIA doesn’t lead the automotive operating systems race, we can’t underestimate the role of chip manufacturers whose GPUs are the core processing engines in many autonomous vehicles. In 2016, NVIDIA introduced DriveWorks Alpha 1, a self-driving car operating system. Even though it didn’t make a big splash in the automotive world, new developments in the framework of this scalable AI platform for autonomous driving appear almost every month. NVIDIA DRIVE is more than just an automotive OS: it’s a huge platform with products and services that will definitely please OEMs and Tier 1 providers.
Automotive brands that use it: DriveWorks is used by 370 of the world’s automakers, Tier 1 suppliers, developers, and researchers including Tesla, VW, Mercedes-Benz, Audi, Veoneer, and Bosch. You can read more about NVIDIA initiatives for connected and autonomous cars here.
Mentor Nucleus® OS
A range of services from Mentor, a Siemens business, help automakers and OEMs meet the growing trend of integrating ADAS, driver information, and infotainment. The Nucleus® real-time autonomous car operating system is one of these services. Plus, Mentor has introduced the DRS360 autonomous driving platform designed to deliver the low-latency, high-accuracy sensing required for Level 5 autonomous vehicles.
You can read more about Mentor initiatives for connected and autonomous cars here.
Linux is a popular OS platform in autonomous cars. Even years ago, the computers in Google’s autonomous car were running Linux, and so were prototypes by VW and GM. There are various versions of Linux, but two of the most widely used in automotive are Ubuntu and embedded Linux.
Like the majority of the automotive operating systems, Linux isn’t a single OS. It’s an entire foundation-based organization called Automotive Grade Linux (AGL) that improves the security of its systems and works on telematics solutions and mapping projects. Beyond the infotainment stack, where Linux is massively used nowadays, real-time car operating system Linux will make a difference at the highest levels of autonomous driving.
Automotive brands that use it: BMW, GM, VW, Toyota, Chevrolet, Honda, Mercedes, Tesla, Lyft, Baidu. You can read more about Linux initiatives for connected and autonomous cars here.
Android Automotive OS
Google isn’t sparing resources to take the lead in managing the infotainment systems of autonomous vehicles. The tech behemoth plans to do this with an extended version of Android Auto, which for now only works when paired with an Android smartphone. Shown on the infotainment display, this connected car operating system provides access to the Google Play ecosystem and connect to system-level operations with an advanced Google Assistant. The system is also customizable, making it even more appealing to automakers and end users.
Automotive brands that use it: Volvo and Audi have signed contracts with Google to start using the upgraded version of this automotive OS in 2020. They’re followed by the Renault-Nissan-Mitsubishi alliance, which is partnering with Google to embed the Android operating system for cars in its future vehicles. You can read more about Google initiatives for connected and autonomous cars here.
If you hear about Google pioneering some kind of service, look out for Apple (and vice versa). True, it’s hard to call Apple’s infotainment offering a full-fledged connected car operating system, but they do have an ace up their sleeve: a so-called “secret” self-driving project named Titan. What isn’t a secret is that Apple is developing its own operating system for autonomous cars produced by this brand as well. Now, all we can do is wait to see how far they’ll extend the current version of CarPlay.
Automotive brands that use it: As of today, Apple CarPlay is offered as a standard or optional feature in numerous vehicles by more than 30 manufacturers including Audi, BMW, Toyota, Nissan, Ford, Honda, Mazda, and Mercedes-Benz.
ROS (Robotic Operating System)
Don’t be deceived by its name: Robotic Operating System isn’t an operating system per se. It’s an open-source software platform and a set of tools for robotics programming. An autonomous vehicle is made up of various ECUs that perform separate functions but should communicate and be managed as a single system. That’s where ROS can help. It works on fully distributed computational terms, meaning that different computers can take part in the control processes but act as a single entity.
The main problem preventing massive adoption of ROS in the automotive industry is security. With no security mechanism to prevent third parties from getting into the ROS network, people will never agree to put their lives in the hands of this open platform. And automakers are perfectly aware of that.
Automotive companies that use it: General Motors, BMW, Ford, Bosch. You can read more about ROS initiatives for connected and autonomous cars here.
Microsoft Windows, the world’s most popular operating system for desktops and laptops, lags when it comes to connected and self-driving initiatives. Microsoft does have a plan, however, to partner with automakers that need an operating system for their vehicles and help them develop their own autonomous driving capabilities. With the Microsoft Connected Vehicle Platform, which includes Azure, Office 365, Cortana, and other cloud-based intelligent services, Microsoft is sure to jump in on the autonomous driving frenzy.
Two approaches to automotive operating systems that automakers and OEMs may pursue
The era of mass implementation of self-driving cars hasn’t arrived yet, so carmakers still have time to select the most rational approach to automotive OS development. Usually, the issue comes down to the proprietary versus open-source question. Companies have to decide whether to continue working on their own independent operating systems or choose a standard operating system to be used across different manufacturers and models.
The first option entails complications like compatability with numerous operating systems, confidentiality agreements, and standardization. The case may get even more dire as non-automotive players continue their expansion into the automotive industry. They’ll definitely respond to the market demand for augmented and virtual reality solutions, virtual assistants, artificial intelligence services, 3D mapping products, and so on.
The second option seems more viable — especially with OEMs, as they often lack the necessary capacity within their organizations to develop their own operating systems. For OEMs, partnership and collaboration fit the bill, but for such a highly competitive industry as automotive, it’s a huge cultural shift. For decades, manufacturers have been striving to maintain control of their branding and technologies, and it won’t be easy to change. Several reasons for that are safety and security challenges.
Regardless of the approach automakers and OEMs choose, their decision will shape the future of automotive software development, and we do hope that the change will be for the better.
You can hardly call the automotive operating systems market established. But it’s growing and getting stronger. The automotive industry has received large injections of money and technology to develop connected car OSs (the first use case of automotive operating systems) and the OS platforms in autonomous cars. Competition between industry players is intensifying: tech giants, OEMs, Tier 1 and Tier 2 providers, and startups are all trying to build the operating system for the car of the future.