. Concurrency faults in software of wireless-sensor-networks are important causes of failures of sensor-network applications. The concurrency errors are triggered by buggy interleavings among interrupt-handling instances. The interleavings among instances at runtime are complicated and non-deterministic. Thus, concurrency faults are always hard to find. This project aims to develop a set of dynamic detection and localization techniques for concurrency bugs, based on bug patterns. Firstly, the project will create a set of dynamic bug patterns to denote problematic interleavings among instances, and develop an instance-identification technique to be the foundation of dynamic analysis of instances. Secondly, the project will develop a localization technique for concurrency faults, and a detection technique for occasional concurrency bugs, respectively. The features of the techniques are: monitoring bug patterns of a tested program at runtime, computing the suspiciousness scores of the patterns according to statistical results, and ranking the bug patterns. The advantages of the techniques are as follows: they can detect multiple types of concurrency bugs, can automatically localize the faults at the level of source-statement, and can visually report the locations, types and causes of the bugs. The successful implementation of this project will contribute to deepening the understanding of the characteristics of concurrency-bugs in sensor-network software, and will provide new research ideas for testing and debugging. Moreover, the tools developed in this project can be applied in practical testing and debugging activities for sensor- network software.
无线传感网软件中的并发错误是导致传感网应用失效的重要原因。这些并发错误由中断处理实例之间的错误交织而造成。因为运行时的实例交织复杂而不确定,所以并发错误往往难以找到。本项目旨在针对并发错误,开发一套自动化的、基于错误模式的动态检测和定位技术。首先,本项目将创建一组动态错误模式以刻画有问题的实例交织,并开发一种实时的实例识别技术以作为动态分析实例的基础。然后,本项目将开发针对并发错误的定位技术以及针对偶然并发错误的检测技术,其技术特点是:在测试运行时监测错误模式,根据统计结果计算模式的可疑度分值,并将错误模式排序。这套技术的优点是:能检测多种类型的并发错误,能自动地定位错误到源码语句级,并能可视地报告错误的位置、类型和原因。该项目的成功实施将有助于深化理解传感网软件并发错误的特征,为测试与调试提供新的研究思路。此外,本项目所开发的工具可应用在实际的传感网软件的测试与调试中。
软件质量是无线传感网得以成功应用的重要保证,而软件缺陷已成为许多无线传感网应用失效的重要原因。本项目研究了无线传感网软件的常见缺陷特征,分析了这些缺陷的原因并建立了一套针对这些缺陷的基准被测程序包。项目深入研究了无线传感网程序中难以被检测的并发错误,建立了一套错误检测模式,包括并发访问内存的错误模式和并发访问信道的错误模式。项目通过中断处理实例之间的错误交织来刻画这些错误模式,并开发了一种实时而高效的中断处理实例识别技术,包括算法、实现、理论证明和实验验证。项目进而开发了一套基于上述并发错误模式的传感网程序测试新技术,能够动态检测和报告并发错误的类型和原因,并能自动化可视化地定位错误到被测程序的源码语句行。项目所开发的测试技术包括算法、原型工具和实验展示。本项目的研究结果能推动传感网软件的测试、调试和质量保证技术的研究,所开发的原型工具可应用在实际TinyOS程序的测试与调试中。
{{i.achievement_title}}
数据更新时间:2023-05-31
路基土水分传感器室内标定方法与影响因素分析
硬件木马:关键问题研究进展及新动向
内质网应激在抗肿瘤治疗中的作用及研究进展
采煤工作面"爆注"一体化防突理论与技术
基于循证实践的腰痛康复治疗国际指南解读与启示
面向高并发软件系统的动态缺陷检测技术研究
基于模型检查的软件错误定位技术研究
面向软件自动修复的错误定位技术研究
软件配置错误的主动检测技术研究