In the automotive industry and the vehicle cockpit multifunction touchscreen displays have replaced the gauges and indicators of the past. High performance SoCs are being used to not only convey basic information to drivers but to achieve augmented vision and improved situational awareness in a human-machine-interface, which allows the driver to visually interpret large amounts of disparate information.
In avionis, GPUs and related software drivers used in these high reliability and safety critical automotive systems will adopt more safety qualified electronics over time. These new systems will be fit for purpose, because they will be designed to meet functional safety standards, enable deterministic computing and software will be written in conformance with the MISRA-C guideline for safety critical software in accordance with safety standards such as ISO 26262.
Workflows: a current trend in automotive is to consolidate multiple cockpit functions into a cockpit domain controller (CDC) where multiple traditional vehicle ECUs can consolidate as software workloads on a single ECU. Often times, workloads that have deterministic compute deadlines may not share resources with other workloads, and this is important for the deterministic safety HMI workloads. The Arm Mali-G78AE Flexible Partitioning feature is an innovative feature that allows boot time HW separation of the GPU into 1, 2, or 4 appropriately sized GPUs. These GPUs can be allocated to the different workloads.
Safe HMI architectures: there are many cases os use for ISO 26262 ASIL B and ASIL ASIL B . The integration of safety content with non-safety content on an instrument cluster is commonplace. In current systems rendering may be performed by ASIL Quality Managed QM hardware and software. The safety mechanism in place is an ASIL B. The most common method is using dedicated hardware in the display controller that generates a CRC. This CRC is compared with a refence CRC value supplied by the application. If they match, you know the QM graphics system is working correctly and the safety related content is displayed correctly. However, this approach does not scale well if this criterion is not met. To support a scalable approach for the future, designs will transition to system architectures that support safety applications running on safety qualified applications cores.
Safe HMI application development: safety applications in avionics take advantage of GPU acceleration enabled by the OpenGL® Safety Critical API. In the future, the VkCore Functional Safety Suite for Mali-G78AE will support OpenGL SC 1.0.1, 2.0 and the Vulkan SC 1.0 API that is under development by the Khronos organization. HMI designers will take advantage of HMI tools that support a workflow allowing for tagging of certain data and/or surfaces for safety and non-safety rendering domains. The ecosystem of HMI tool partners will support workflows for an easy transition from ASIL QM HMIs to Safe HMIs that are composited safely with non-safety HMIs.
Strong Partnerships: as automotive systems become more autonomous and as artificial intelligence and machine learning becomes more pervasive, technology partnerships in the automotive world will become more commonplace.