Automatic Test Case Generation for Smart Human-Centric Ecosystems

Smart Human-Centric Ecosystems (SHcES) are an important class of systems that impact on our everyday life. They span from smart homes to smart cities, smart grids, autonomous vehicles, smart schools, and smart healthcare systems. A SHcES comprises a wide range of autonomous and heterogeneous hardware and software systems that interact explicitly and implicitly in a shared environment. The core elements of SHcES are humans. Humans interact with the systems within the SHcES both explicitly through system interfaces and implicitly by freely acting in the ecosystem. For instance, humans interact with many systems in a smart city even by simply standing or moving. Classic testing approaches assume that users interact with the system only through the system interfaces, and thus miss many relevant elements of the sequences of human actions that comprise test cases of SHcES. The complete freedom of human actions and the vast variety of interactions that characterize the human behavior in SHcES make it challenging to generate sequences of human actions that constitute the backbone of test cases for SHcES. We hypothesize that sequences of human actions in SHcES depend on the personality, and this relation helps us automatically generate complete sequences of human actions that comprise test cases for SHcES. We assume that it is possible to infer the personality of humans by observing the human actions in the SHcES up to a given instant and use this information to infer the most likely human actions that can follow. We use this information to automatically generate test sequences for SHcES. With the use of personality models, we can inject new personalities in the SHcES, to test how different personalities and social groups interact in the context of new scenarios and system behavior before they happen in the field.