Analysis of real-time scheduling algorithms under overload conditions in a multi-core ECU

As a vehicle’s software functionality increases, the number of ECUs required also increases. The addition of more ECUs to a vehicle is not a beneficial modification from a multitude of perspectives, including financial, operational, and environmental. One possible method for reducing the number of ECUs is to consolidate them using multi-core processors. In this paper, we apply representative real-time scheduling algorithms to a multi-core processor and analyze the stability of each algorithm when the ECU is overloaded. We also evaluate two methods for distributing tasks from a communication core to processing cores. One is based on odd and even priorities, and the other is a random, priority-independent distribution.