Generating Test Suites for an Embedded ESP8266 System by Using Black-Box Testing Methods

Black-box testing methods are suitable for application on embedded systems due to the lack of usage of programming code and simplistic approach of comparing expected and actual outputs on live or simulated system components. There is a low number of studies that explain the process of black-box testing in detail or that apply multiple methods and compare the structures of the generated test suites. This paper proposes an embedded system programmed on the ESP8266 microcontroller chip for the visualization of environmental sound by using a sound detection sensor and a NeoPixel LED ring. The visualization is performed by using both LED color and number of LEDs that are turned on. Three different black-box testing methods were applied for generating test suites for the proposed system: equivalence partitioning, pairwise testing and cause-effect graphing (Myers’ back-propagation algorithm and forward-propagation algorithm). Equivalence partitioning and Myers’ algorithm yielded redundant test cases. Equivalence partitioning test suite had a considerably lower number of test cases compared to the other algorithms. An overall overlap of 23.33% of maximum unique test suite size was achieved as a result, whereas the large test suites had a high overlap percentage. This indicates that future integrations between different black-box testing methods may be possible.