A Task Co-Scheduling Research On Mixed-Criticality Automotive Electronic Systems

Modern automotive electronic systems are heterogeneous distributed mixed-criticality embedded systems with multi-functions integrated.Efficient scheduling of multiple functions affects the use of system resources.Functional safety is one of the important research directions in this field.Meeting the real-time and reliability requirement of the function is an important requirement for ensuring functional safety.However,for automotive resourcesconstrained embedded systems,it is difficult to meet the real-time requirements of all functions.The real-time of automotive electronic systems depends on the real-time of highcriticality functions.Meanwhile,as industrialized products,hardware cost is also an important cost factor for automotive electronic systems to be considered during the design phase.Therefore,this paper focuses on modern automotive electronic systems and study automotive mixed-criticality task co-scheduling towards on safety,reliability and hardware cost optimization.The algorithm is designed from the perspective of optimization of functional reliability,real-time performance of high-critical functions and hardware cost constraints.The main works are as follows:An algorithm named Multifunctional Reliability-aware Efficient Scheduling(MRES)is proposed for response time optimization under reliability requirement assurance of mixedcriticality functions.It improves the optimization trade-off between reliability and real-time performance of mixed-criticality functions,also covers the shortage of current research which merely considering single-function multi-objective optimization or multi-function single-objective optimization.The problem is decomposed into two sub-problems,namely,meeting reliability requirements and response time optimization.The proposed algorithm realize the transfer of reliability requirement of function to that of each task by a preallocation approach,and the ECU that does not meet the reliability requirement is filtered during task assignment;It assign tasks according to the minimum EFT policy and re-allocate ECU resources by considering the criticality of functions.Experiments on the small-scale and large-scale automotive functions show that the proposed algorithm can reduce the DMR of high-criticality functions more effectively than the existing algorithms under reliability requirements assurance of all functions.An algorithm named Fault-Tolerant Hardware Cost-aware Multifunctional Safety Assurance(FT HCMSA)is proposed for response time optimization under reliability requirement assurance and hardware cost optimization of mixed-criticality functions.It improves the optimization trade-off between hardware cost constraint and multi-objective scheduling,and achieves the goal of considering functional reliability and real-time optimization under hardware cost constraints.The FT HCMSA algorithm uses task fault tolerance and uses the replacement and removal strategies to optimize hardware costs based on the MRES algorithm proposed in this paper;It defines a task reliability requirement under task fault tolerance and propose a more granular task allocation strategy.Experiments on the reallife automotive function and synthetic randomly generated automotive functions show that the proposed FT HCMSA algorithm can reduce the response time of high-criticality functions more effectively than the existing algorithms under the constraints of functional reliability and hardware cost.