个人信息Personal Information
教授
博士生导师
硕士生导师
任职 : 大连理工大学软件评测中心主任
性别:男
毕业院校:哈尔滨工业大学
学位:博士
所在单位:软件学院、国际信息与软件学院
学科:软件工程. 计算机系统结构
办公地点:开发区校区综合楼409
联系方式:zhoukj@dlut.edu.cn 13804248599
电子邮箱:zhoukj@dlut.edu.cn
Automated Software Testing Based on Hierarchical State Transition Matrix for Smart TV
点击次数:
论文类型:期刊论文
发表时间:2017-01-01
发表刊物:IEEE ACCESS
收录刊物:SCIE、EI
卷号:5
页面范围:6492-6501
ISSN号:2169-3536
关键字:Hierarchical state transition matrix; Cyclomatic complexity; expanded regular expressions; group state
摘要:Smart TV in China as an important component of the smart home, does not only have the functions of the traditional TV, but also have the functions, such as distance education, remote monitoring, E-business, and media playing, which brings about its software to be more complex in structure and larger in scale, accordingly, the total testing efficiency becomes lower when using traditional testing methods, and the deep-hidden software defects cannot be detected efficiently and effectively. A novel automatic software testing method based on system design specifications is proposed to improve the smart TV software testing efficiency. First, the behavior of the smart TV is modeled, based on the system design specification with hierarchical state transition matrixes (HSTMs). The scale of the state model of the smart TV is lowered by setting the group state according to the choice of the key nodes based on the importance of the nodes in the network; then, the HSTM model is converted into an expanded regular expression (ERE) with the memory property. Second, every closure operator in the ERE is replaced recursively with a certain integral value, according to the cyclomatic complexity of an ERE in the closure to generate a simplified ERE. Then, a test case is generated from the simplified ERE. Finally, the test cases are converted into python script, and a test platform is designed to send the python script to the Android smart TV automatically through its android debug bridge interface. The practical application shows that the test period is shortened, and comparing with the traditional manual test methods, more errors can be tested.