Research On Energy And Cost Sensitive Scheduling In Automotive Cyber-physical System

Modern automotive electronic systems have evolved into automotive cyber physical systems,namely ACPS(Automotive Cyber-Physical Systems).It is a heterogeneous distributed embedded system with deep integration of computing,networking and physical world.ACPS is characterized by its heterogeneity and dynamics.At the same time,new energy automotive is the inevitable path for the automotive industry to adhere to sustainable green environmental development.Limited energy resources are also an important feature of ACPS.How to optimize the real-time,energy consumption and cost resource consumption in the heterogeneity and dynamics of ACPS is a new research hotspot.From the perspective of energy consumption and cost resource sensitivity,this paper designs a scheduling algorithm from the perspective of optimization of functional applications security,energy resource consumption and cost resource consumption for parallel distributed function applications in ACPS with limited system resources.To ensure the safe scheduling,low energy consumption and low cost of dynamic function applications in ACPS.The main work of this paper is as follows:Aiming to the real-time and energy-optimization problems of dynamic functions in ACPS,a collaborative optimization mechanism for functional application security and energy consumption is proposed from a global perspective.The basic idea is: from the perspective of functional application security,a deadline-driven processor merge algorithm for multiple dynamic functional applications is proposed.From a global perspective,the basic problem of maximizing the number of functions completed by the dynamic arrival function within the deadline is realized;In addition,from the perspective of energy optimization,a global energy-saving scheduling algorithm is proposed,and combined with the characteristics of the ACPS integration architecture,the energy consumption optimization of the functional application that meets the deadline is realized by selecting multiple candidate executable ECUs for the execution task.This article uses a real parallel function application to conduct a comparative experiment.The experimental results show that compared with the existing methods,the proposed combination of the two algorithms can effectively reduce the deadline missed radio(Deadline Missed Ratio,DMR),and its reducing the actual energy consumption of the system as much as possible.The optimization of cost resources under the condition of energy and response time constraints in ACPS is studied.In order to simplify the problem,a two-stage solution is proposed.The first stage is to solve the response time optimization problem under energy constraints,and an energy-constrained response time optimization algorithm based on a more aggressive energy pre-allocation strategy is proposed.The algorithm uses a preallocation method to pre-allocate unallocated tasks,and then converts the energy constraints of the functions to the energy constraints of the tasks;The second stage is to solve the cost optimization problem under the condition of energy and response time constraints on the basis of the first stage.Based on the global strategy,an efficient energy and response time constrained cost optimization algorithm is proposed.The algorithm performs energy redistribution based on the first stage energy pre-allocation,and utilizes the relaxation time available on the processor to allocate the lowest cost processor to the task under the constraint of energy and response time.The proposed algorithm is validated using real automotive function applications and randomly generated functional applications,The experimental results show that compared with the comparison algorithm,the proposed solution not only ensures the constraint of energy and response time,but also can greatly reduce system cost resource consumption.