Automation of Software and Hardware Systems Acceptance Testing in the Paradigm of Behavior-Driven Development

The paper is devoted to the problem of software and hardware systems acceptance testing according to the Agile Testing methodology. The Agile approach is widely used by software developers, however, developers of software and hardware solutions rarely use this approach, not believing in its effectiveness. The paper assumes that for software and hardware complexes, the practice of continuous integration should be applicable and, as a result, test interaction with hardware at the level ofphysical interfaces performs automatically. The article's authors presented Accepta, a system for automating acceptance testing of software and hardware complexes, explicitly designed for the Agile Testing methodology in the context of continuous integration. The main component of Accepta is an interface block based on the Nucleo-F767ZI debug board manufactured by ST. Test actions that run within the framework of Accepta supplement by expanding the command system of the interface block. The software part, which implements the functions of test describing and execution, is based on the Cucumber framework for automating software systems acceptance testing. The requirements for the object under test and the scenarios for checking the requirements are described in the Gherkin language, which is close to the natural description. The test script steps are described programmatically in the Ruby language. The actual execution of test actions is provided by sending commands through the COM port to the interface unit and analyzing the received responses. As the practice of using Accepta in working projects has shown, this approach allows us to successfully apply the Agile development methodology for software and hardware systems. Due to the automotive interaction with the device under the test interface, high intensity of testing in the development process, including regression, is ensured. The regular testing consequence is fast feedback: as soon as any functionality stops working correctly, developers find out about it fast. At the same time, due to the use of test automation tools, the reproducibility of test action sequences led to the detection of a defect is ensured.