Research On Intelligent Deployment And Reliability Algorithm Of Automotive Electronic Software

Recently automotive electronic technology has been widely used in all areas of the automotive industry.It provides important technical support for safety,environmental protection,energy-saving,intellectualization and informatization of vehicle.To meet the market requirements,the hardware architecture of automotive electronic control unit(ECU)becomes complex and the number of software applications is so huge and with wide varieties.It puts forward a new challenge to the development of automotive electronic software and system integration.The automotive electronic software system belongs to safety-critical areas.It not only requires the functional correctness of the system,but also has very high requirements for the real-time and reliability of the system.The automotive electronic system has its own unique hardware and software architecture,such as the distributed hardware topology,the applications composed of the components,the runnable entities with complex temporal property and etc.Therefore,the automotive electronic system is limited by distributed structure,bus load,computing capability of the electronic unit and so on.In the process of automotive electronic software system integration,it is necessary to not only satisfy the real-time requirement of software components under resource constraints,but also to evaluate the reliability of the system.This presents another challenge to the automotive electronic system integration.In the dissertation,the hardware model and software model based on the hardware topology and the software archetecture are specified formally.Then,the automotive electronic software deployment problem in AUTOSAR architecture is studied.Finally,in order to satisfy the system for the requirement of real-time and running efficiency,several software integration and optimization methods in automotive electronic distributed system are proposed and a new method for evaluating the reliability of the system is proposed.The main work and innovations of this dissertation are as follows:1.The sortware component mapping problem in automotive electronic distributed system is studied and a now intelligent mapping method base on clustering algrothm is proposed.Our method designs a nonlinear adaptive function to optimize the bus traffic and the load of each ECU.The function adjusts the weight of the double optimization target in the clustering process dynamically and ensures the system schedulability by adjusting its parameter and threshold value.The experimental results show that our method not only reduces the bus traffic,but also optimizes the load of each ECU,so that the system load reaches equilibrium between ECUs.2.The scheduling strategy of operating system in AUTOSAR standard is studied and a mapping method from the runnable entities in software component to operating system tasks is proposed.Based on hierarchical clustering algorithm framework,our method designs and implements a filtering and weighting table which makes the jitter time of task,the blocking time between tasks,the scheduling frequency of system and the data traffic between tasks are considered in clustering process.The experimental results show that our method eliminates the task jitter time,decreases the task blocking time,system scheduling frequency and the data traffic between tasks.3.The task scheduling problem under multi-core ECU is studied and two task assignment methods are proposed specific to the general scheduling strategy and the scheduling strategy based on AUTOSAR dispatcher respectively.The first method searches for the current optimal ECU core for each task based on the task's worst response time.The second method computes the optimal activation offset in dispatcher for each runnable entity and then calculates the optimal allocation strategy in multi-core ECU for each task.The second method enhances the time determination of each runnable entity during its running time.The experimental results show that the execution time of the runnable entities is uniformly distributed in the time axis of each ECU core in the second method,so that the task's response time in the second method is shorter than that of the first.4.The fault tolerance mechanism and error management mechanism in automotive electronic system is studied and a software system reliability model in the design phase is established and a reliability evaluation method based on error scenario is proposed.Our method creates the error scenario for the system based on the reliability model,improves the reliability calculation method to speed up the process of reliability calculation.The experimental results show that our method not only calculates the system reliability,but also guides the software design from the aspect of reliability.