Research On Status Monitoring And Self-recovery Technology For Embedded Software

With the rapid development of computer technology and microelectronics technology,the application of embedded system has been spread to every corner of the society. Theincreasing complexity of embedded software, the hidden troubles of the embedded system aremore and more. Once the system fault or software failure in the process of running forembedded software, it will be difficult to complete the mission, there may even cause hugedamage to property. Therefore, the reliability of embedded software has become a hotresearch direction for the moment. Taking the embedded software of VxWorks as the researchobject, this article conducts a research on the embedded software and self-recoverytechnology. The main work is as follows:Firstly, this article expounds the research status at home and abroad for embeddedsoftware status monitoring and self-recovery technology, briefly introduces the researchbackground and significance of this topic, and shows the content of the main research; Mainlyintroduced the related embedded software self-recovery technology, including theintroduction of the necessity of the embedded software self-recovery and the main problemsin the process of research; Respectively introduced about the two common embeddedsoftware self-recovery strategy as well as their advantages and disadvantages; Described indetail the most commonly used repair action in the self-recovery strategy--microreboottechnology.Secondly, research on the method of embedded software status monitoring from twoaspects of the embedded system performance and embedded software behavior respectively.In the study of the system performance monitoring method, four indicators carried outin-depth research which including the system memory, the running status of the task itself,and the occupancy of both CPU and task stack; At the same time, a method of embeddedsoftware aging trend analysis was proposed, which uses the method of least squares for curvefitting and adopts a threshold-based embedded software failure time estimation method. In thestudy of the software behavior monitoring method, using a method of building softwarebehavior tree for software behavior monitoring, and introduces the classification of nodes,themethod of building,and the matching rules of behavior in the software behavior tree.Again, combined with the characteristics of the VxWorks operating system, a nested self-recovery strategy for embedded software was proposed; respectively introduced theembedded software self-recovery strategy based on system performance monitoring and thedetail process of embedded software self-recovery strategy based on software behaviormonitoring. Based on the result of embedded software status monitoring in the third chapter,using the selected detection indicators to build the state space; According to the selectedindicators of abnormal state of corresponding repair behavior to build action space, and givenrepair action evaluation model. On the basis, using a rule-based embedded softwareself-recovery decision-making method.Finally, for the proposed method of embedded software status monitoring and embeddedsoftware self-recovery strategy was verified by experiment, the results show that thesemethods are feasible and effective; By comparison, the superiority of nested self-recoverystrategy is proved.