A major challenge facing automobile manufacturers is the sheer number of ADAS-related electronic control units (ECUs) found in the car: distance alert, lane assist, rear-view camera, surround view, parking assist, and the list gets longer with each new model. This affects the weight of the car and therefore has an impact on its energy consumption. Adding a great amount of hardware to a vehicle has a negative impact both environmentally and financially.
COQOS Hypervisor SDK is the practical solution to integrate multiple functions on a single ECU. The COQOS hypervisor creates separated Virtual Machines (VMs) where the customer can integrate operating systems with different requirements and safety levels (ASIL). Thus, ADAS systems from different manufacturers can be executed on one processor.
The integration of functions into different VMs also allows for flexible allocation to the processor cores of the multi-core processors: Guest operating systems with high computing power requirements can gain access to multiple cores. Conversely, multiple VMs can share a single kernel deterministically. Through this dynamic resource allocation, COQOS Hypervisor SDK ensures optimal utilization of processor processing power.
ADAS systems based on COQOS Hypervisor SDK can be integrated into the cockpit controller. For example, the infotainment system runs in one of the virtual machines created by COQOS Hypervisor SDK, while the software of the rear-view camera is integrated into a second VM. Both systems are running on one processor. This means that the hardware resources of an ECU are split between the infotainment functions and the ADAS. This allocation of computational power will be particularly interesting when using the latest generation processors that provide enough performance for highly complex features. The performance of these systems is crucial for them to function properly. This includes everything from fast boot and real-time to smooth images from the rearview cameras. In the long term, this development leads to automated driving and on to autonomous driving.