Centralizing ADAS- A Look into the Automotive Nervous System

I grew up in India and my first car was an old white 80’s Suzuki 800 cc that belonged to my parents. It was very basic and didn’t even have a radio because my father thought it was a distraction. When I finally convinced him to add a radio, it was underpowered, prone to breaking down, and very unattractive; but it was my first car, and I loved it. Today, my car and the majority of cars built in the last few years are a lot more intelligent. They know if you put your seatbelt on, can tell if there is someone in the lane next to you, and can even brake automatically in response to changing road conditions ahead. All of these features are part of something collectively referred to as Advanced Driver Assistance Systems (ADAS). This is a term for a variety of technologies that are designed to help a car be safer by alerting the driver of potential hazards or by automatically taking over control of certain critical functions from the driver.

Automatic braking and lane assistance are great features that help protect the driver and passengers. Although these systems all work on keeping us safe, these early ADAS solutions are distinct systems that operate separately from each other. Much like early computers, they are function specific with each system analyzing its own data set and performing its own task. If these systems were to be centralized, combining the data, more advanced safety protocols could be implemented. By uniting systems like automatic braking and lane assistance, a centralized ADAS system would detect traffic build up and look for an open lane next to the vehicle. Simultaneously, automatic braking would activate the turn signal, informing the driver that there is an open lane, and warning drivers nearby about the impending lane change. This second degree of freedom (to move to the left or right, instead of only having the option to slow down), if implemented correctly, would give a car more options to avoid an accident.

The example above is simple, but illustrates the potential benefits of a centralized ADAS. Instead of a number of discrete systems all operating independently, a centralized system would have access to all of the sensors within each system, and process all of that data at a central location. In addition to enabling a number of new use cases that would help make ADAS more effective, it would also be more cost effective. As ADAS evolves into semi-automated and fully automated solutions, centralization is a logical step. And according to a recent ABI Research report, NVIDIA, NXP, and Mobileye, all major players in the existing ADAS space, have already announced centralized automation driving platforms. These solutions may form the platforms upon which the next generation of automation is built.

In order to do this sort of integration, standards would have to be created that allow the ADAS components to communicate according to established and agreed upon norms. Much of this will have to be done responsively, as market forces produce winning solutions that are then codified. However, given the variety of sensors, standards and technologies in play, even with such standards established, some degree of flexibility will still be required. Automotive FPGAs will help bridge the gap as hardware from a variety of manufacturers and interface protocols communicate together. In addition, FPGAs will assist the processors with performing and assisting in functions, improving overall performance and response time.

Centralizing control and processing is a common theme in technology. Market forces demand that after best practices are discovered, integration occurs to lower the cost of implementing such solutions. But technology is also about continuous innovation. In ADAS, a reliable centralizing automation is not only efficient, it is necessary to unlocking the full potential of the automated automotive future.