Dynamic Task Allocation and Scheduling for Multicore Electronics Control Unit (ECU)

In recent years, an observable trend in the automotive industry shows that more and more electronic features are getting added up. Hence the demand of computational resources in the vehicle is growing rapidly. A good scheduling strategy needed to utilize the limited resources of a multicore system to give a better performance in terms of preemption, migration, deadline-miss and dedicated resources for safety critical tasks. In this paper, a Dynamic Task Allocation and Scheduling for Multicore Electronic Control Unit is proposed. The focus is on engine control unit (ECU) where tasks execution changes dynamically as per engine speed. In the tri-core scheduling, one core is reserved as dedicated core for safety and critical task execution and remaining two cores are used for non-critical tasks. Proposed scheduling technique is a hybrid scheduler which is a combination of least laxity first (LLF) and earliest deadline first (EDF) schedulers. The extensive simulations carried out on hybrid scheduler shows better performance in terms of increased core utilization, less context switching and guarantee to meet deadline compared to the existing dynamic schedulers.