In order to improve the testing efficiency of embedded software,model-driven embedded software testing has become a research hotspot in recent years.It uses the model to describe the software functions,effectively solving the problem that the embedded software is limited by hardware environment.However,due to the complicated technical process,we found that the current testing technology has performance problems.When conducting large-scale comprehensive testing,it takes a long time to execute all the test cases,and the software testing efficiency is not well.Due to the particularity of the embedded software,a virtual simulation environment needs to be provided during the testing process.However,the existing virtual simulation technology can not guarantee the efficient operation of real-time software,which affects the efficiency of software virtual execution.In order to solve the above problems,this dissertation deeply studied the embedded software automation testing technology.The specific work is as follows:(1)In order to improve the testing ability of test tools under heavy tasks and shorten the time for performing large-scale comprehensive testing,this dissertation conducts in-depth research on cluster architecture and proposes test cluster execution technology based on use case importance.Firstly,because important test cases are prioritized to detect errors as early as possible,the test cases are sorted by the optimized test case prioritization algorithm to ensure that the higher priority test cases can be executed immediatly,which can guarantee the test efficiency and make sure software bugs are discovered as early as possible;then,the sorted test cases are transferred to the cluster environment.Considering the performance and real-time load of the test server cluster,we use consistent load-based dynamic load balancing strategy to evaluate the test cases in a fixed period,balance the distributed test cases to the cluster,improve the throughput of the entire cluster,and improve the efficiency of software testing.(2)In order to solve the problem of inefficient execution of embedded software in virtual simulation environment,this dissertation studies the embedded virtual simulation technology used by the test server in the cluster to improve the performance of virtual simulation environment.Based on hardware virtualization technology,we have improved memory virtualization and scheduling strategies technology,effectively solve the defects of traditional methods.It is designed to reduce the performance loss caused by host and virtual machine switching operation,and ensure that the processor resources occupied by virtual machines can not deprived during the whole process of embedded software testing,which improves the efficiency of embedded virtual simulation.(3)In order to apply the above research results,this dissertation is designed and implemented an efficient automated test platform for embedded software based on virtual simulation,and verified the test platform by the embedded software example of UAV flight control system.The experimental results fully demonstrated the effectiveness of the proposed method. |