AutoMOTGen: Automatic Model Oriented Test Generator for Embedded Control Systems
AutoMOTGen: Automatic Model Oriented Test Generator for Embedded Control Systems
复制标题
AutoMOTGen：面向嵌入式控制系统的自动模型测试生成器
DOI：
--
复制
发表时间：
2008
期刊：
影响因子：
--
通讯作者：
K. Shashidhar
中科院分区：
文献类型：
--
作者：
Ambar A. Gadkari;Anand Yeolekar;J. Suresh;S. Ramesh;S. Mohalik;K. Shashidhar
我们提出AutoMOTGen，一个用于自动测试用例生成（ATG）的工具，用于测试汽车控制器的MATLAB Simulink/Stateflow（SL/SF）模型[6]。我们的方法基于模型检查[2]。该工具的主要亮点是：1增强了模型元素和高级需求的覆盖范围。
1采用模块化设计，即插即用的不同的模型检查器，测试数据生成器和覆盖率分析工具，以提高测试套件的质量。
1实现采样时间抽象，以在中间模型中使用较少的（离散）步骤生成测试。
1为结构覆盖标准实现模型的覆盖相关插装。
1能够处理汽车控制器中常用的SL/SF模块（包括积分器、延迟、乘法/除法、查找表、触发子系统以及分层和并行图表等模块）。
AutoMOTGen的当前实现使用SAL [8]作为中间表示，并使用相关工具（如sal-atg，sal-bmc和sal-smc）生成测试数据并证明某些覆盖目标的不可达性。AutoMOTGen在Java和C++（.NET框架）中实现，并使用MATLAB脚本语言从SL/SF模型中提取测试生成所需的相关信息。
We present AutoMOTGen, a tool for automatic test case generation (ATG) from MATLAB Simulink/Stateflow (SL/SF) models [6] for testing automotive controllers. Our methodology is based on model checking [2]. The main highlights of the tool are: 1 Enhanced coverage of the model elements as well as high-level requirements.
1 A modular design for plug-and-playof different model checkers, test data generators and coverage analysis tools for enhancing the test suite quality.
1 Implements sampling time abstraction to generate tests with lessernumber of (discrete) steps in the intermediate model.
1 Implements coverage dependent instrumentation of the model for the structural coverage criteria.
1 Capability to handle SL/SF blocks commonly used in automotive controllers (including blocks such as integrator, delay, multiplication/division, look-up tables, triggered subsystems and hierarchical and parallel charts).
The current implementation of AutoMOTGen uses SAL [8] as an intermediate representation and uses associated tools such as sal-atg , sal-bmc and sal-smc for generation of test data and proving the unreachability of some of the coverage goals. AutoMOTGen is implemented in Java and C++ (.NET framework) and uses MATLAB scripting language for extracting the relevant information from SL/SF models required for the purpose of test generation.