How is Vehicle Network System Catching up with the Ever-Evolving Automotive Electronics
The networking needs of automobiles have diversified. A safety feature or a drive-by-wire system requires a fault-tolerant system along with time-triggered protocols. Features like Anti-lock braking system, Telematics and FOTA require higher bandwidth. On the other hand, a slower bandwidth like LIN will be sufficient for a few body control features such as wiper control. Add to them an ever-increasing number of software features that lead to more ECUs and more software written on them!
The days of point-to-point connections are long over. Ever since CAN protocol has taken over the inter-ECU communication, the wiring complexity has significantly reduced. However, the pertinent question is- ‘Is the current on-board network architecture with distributed computing good enough?’ The answer to this question is a simple ‘no’.
Where does Current On-board Network System Fall Short?
We talked about the increasing number of ECUs in the vehicle. Along with this sharp increase, there has been continuous demands for better fuel efficiency, functional safety, ADAS features for drivers’ comfort and more.
Such advanced systems need data to move very fast (in the tune of 100 Mbps) between the ECUs and even to the external systems (V2X). Existing ECU communication protocols such as CAN, FlexRay and LIN are here to stay but induction of newer protocols such as Automotive Ethernet has also begun.
As new protocols enter the fray, existing network architecture will need to be updated. For instance, introduction of Ethernet will create a need for middleware solution. Along with that, an increased risk of cybersecurity will warrant additional security measures to be implemented.
In a nutshell, the existing network architecture will need to be revamped. And that brings us to the most important question. What will the future on-board vehicle network system look like? Let’s try to find the answer!
What does Future On-board Vehicle Network Architecture Look Like?
Considering the technological advancements and mapping these with the evolving requirements of the automotive industry, we can classify the evolution of vehicle network into three categories- Upcoming, Future and Future Plus. Let’s discuss each one of them.
The Upcoming network architecture- Domain Fusion: The automotive ECUs are categorized based on their domains and assigned under specific domain controllers. Depending on the vehicle variants, there are domain controllers such as ADAS Controller, Cockpit Controller, Powertrain, and Body Controller. The upcoming network architecture will see the fusion of these domains over Automotive Ethernet network.
Future of vehicle network- Zonal Gateways: Ditch the ECUs and switch to sensors and actuators! That’s where vehicle network is headed to. The sensors and actuators replace the vehicle ECUs and are connected to a central computing platform, kind of an all-in-one processing unit. The sensors and actuators will still use legacy networks such as CAN, CAN FD and LIN and will be connected to zonal gateways which in turn use Ethernet for communication.
How does replacing ECUs with Sensors/Actuator (S/A) help, you may ask? ECUs also have a set of S/As that feeds vehicle information for processing. ECUs are computing unit in itself and can make decisions. Higher number of ECUs comes with its own drawbacks that cause complexity. Having one central computing unit does away with this complexity. Zonal gateways are still required as variety of topologies are used in a vehicle network.
The Future + of vehicle architecture: Centralized Computing: A world without ECUs! Yes, the future of automotive networking completely ditches the ECUs as well as the legacy protocols such as CAN, LIN, MOST and FlexRay etc. Its all about Ethernet now. All the different sensors and actuators are directly connected to a central computing platform which processes the information and control all aspect of vehicle electronics. Since there is no other communication medium, the need for a gateway is also eliminated.
Conclusion
The future of vehicle network is all about data speed and computing prowess, the pre-requisites of almost every innovation in automotive electronics. However, cost-effectiveness is equally important from the business viability standpoint. As far as the future network architectures are concerned, the transition is not going to be very quick. The OEMs are not going to throw away the legacy ECUs that have been around for decades and adopt sensors and actuators. Going forward, we could see a mix of vehicle network architectures in future which will be based on specific needs. Middleware solutions like SOME/IP and gateways would ensure the compatibility.
