Intelligence and net connection have become the major trend in the current development of automobiles. Classic AutoSAR is a software architecture developed on strong real-time embedded OS. The system can meet the customized functional requirements of traditional automobiles; however, once the car is connected to the network, delays and interferences will occur, which hinder real time communication.
In this case, Adaptive AUTOSAR is born to fulfill the non-real-time architecture system software requirements. It coexists with Classic AutoSAR for a long time. Since the safety level of Adaptive AUTOSAR is only ASIL-B (the highest is ASIL-D), Classic AutoSAR (which can meet ASIL-D) is still required for a stronger safety level ECU (Electronic Control Units).
Classic Autosar VS Adaptive Autosar
Current automotive controllers such as ECUs are distinctly different from other functional or infotainment controllers. Controllers based on the Classic AutoSAR platform have the following characteristics:
Infotainment controllers are the opposite as they generally take up larger hardware resources and are not real-time. As they generally run-on embedded PCs, such as LINUX, rather than on automotive-grade operating systems, their failure usually will not cause serious safety incidents. While Adaptive AutoSAR, as a software architecture for heterogeneous software platforms, is mainly used for domain controllers. It can be a bridge between Classic AutoSAR and non-real-time OS like Linux, which has the following features:
Compared with Classic AutoSAR, Adaptive AutoSAR has reduced real-time requirements but greatly improved the support for high-performance processing power, to support the development of intelligent connected application functions while ensuring a certain level of functional safety at the same time. Therefore, C++ will become the main development language for the Adaptive AutoSAR platform.
Software components executing on microprocessors are typically based on the standard of Adaptive AutoSAR instead of Classic AutoSAR. To meet the needs of modularity, dynamism, and continuous update capability, Adaptive AutoSAR has become the agreed software standard on automotive high-performance platforms. It utilizes POSIX-compliant operating systems such as Linux, PikeOS, or QNX, and provides functional extensions for its applications in automotive industry. One of the main function of Adaptive AutoSAR is the communication layer ara::com, which is responsible for communication between Adaptive AutoSAR applications and with other ECUs as well.
Despite certain differences in design and technology ara::com, is comparable to the classic AutoSAR’s functions which includes the middleware that allows for a service-oriented approach. The classic AutoSAR’s software is fundamental and the service-oriented approach allows for the development of dynamic communication pathways during the run time. This process is critical for the software installation while the ECU is running.
In contrast, in Classic AutoSAR, the communication matrix must be modified before the ECU can receive or send new content. By utilizing a service-oriented approach, dynamic subscriptions can be created, leading to a more distinct separation of hardware drivers and upper-layer software. This makes hardware-independent applications in cars highly portable, resulting in even more optimized resource allocation than in the ECU based on Classic AutoSAR. For instance, during the development phase, if the resources of a specific ECU reach the maximum, the software can be easily transferred to another ECU or multiple ECUs without modifying the hardware. This greatly enhances the reusability of software components across different models.
The segregation of software and hardware in the Adaptive AutoSAR project has led to a new distribution of tasks between OEMs and suppliers. A functional block was usually ordered as a physical device in the vehicle previously. It is now possible to purchase only the software. By installing apps from the app store, the driver can become a software integrator himself.
Who is responsible for the failure of the system? Untested application combinations can be installed in the vehicle, which conflicts with the traditional ECU integration requirements. Every configuration under Classic AutoSAR needs to be thoroughly tested. To avoid testing all combinations of applications, it is necessary to ensure there’s no interference between each application. The operating system can ensure that security-related applications do not exceed the memory limit. To achieve this, the operating system provides hard real-time scheduling methods that can define memory limits and worst-case execution times for the applications. Load variations due to interruptions become irrelevant as there is no direct interaction with the hardware.
Adaptive AutoSAR is not designed to replace the Classic AutoSAR platform, but to achieve the coexistence and collaboration of these two applications. Classic AutoSAR platform supports high security and high real-time application scenarios, so embedded software functions need to be deployed and run on the Classic platform; while Adaptive Autosar on the other hand supports parallel processing of big data, so high-performance computing functions need to run on the Adaptive platform.
With the rapid development of autonomous driving technology, along with IoT and cloud technology, Adaptive AutoSAR is emerging to meet not only existing needs but also the revolutionary changes in future automotive technology. With different types of architectures and complements, it can support various adaptive deployments, complex microcontrollers, and the interaction of various non-AutoSAR systems.
Our hardware (VCU and ADCU) is fully capable of AutoSAR, and we have stable connections with AutoSAR provider. Ecotron is focused on the next-generation automotive basic software architectures that will be the standard options for automakers, suppliers, and software providers. We are committed to providing the highest quality products and because we know AutoSAR is a standard option, we are fully capable of applying it to our VCUs. Please email us to find out what we can offer.”
For Business inquiries, please email info@ecotron.ai