Narrow Results

There has been intensive research in memoryless nonlinear behavioral modeling of power amplifiers (PAs). But in broadband communication systems, memory effects of PAs can no longer be ignored and traditional memoryless model cannot accurately characterize the input-output relationship of PAs. In order to treat memory effects and reduce the complexity of general Volterra model, a new behavioral PA model based on modified Volterra series is proposed. Since the characteristics of power amplifiers change during transmission time, a recursive least squares algorithm with size-fixed observation matrices is developed to update the parameters of the PA model. This identification algorithm, which uses only the latest sample data to identify the parameters, can decrease computational complexity and data storage space needed for identification. Simulations are carried out to validate the performances of the proposed PA model and identification algorithm.

UML is now popularly applied as a requirements modeling language for software system analysis and design, and the dynamic behaviors of system are described in UML behavioral model. As the UML model suffers from lack of well-defined formal semantics, it is difficult to formally analyze and verify the behavioral model. The paper presents a method of UML behavioral model verification based on Description Logic system and its formal inference. The semantics of UML behavioral models is divided into static semantics and dynamic semantics, which are formally specified in OWL DL ontology and DL-Safe rules. To check the consistency of the behavioral models, the algorithms are provided for transforming UML behavioral models into OWL DL ontology, and hence model consistency can be verified through formal reasoning with a DL supporting reasoner Pellet. A case study is provided to demonstrate applicability of the method.

Automated GUI testing based on behavioral model is one of the most efficient testing approaches. By mining user usage, test scenarios can be generated based on statistical models such as Markov chain. However, these works require static analysis before starting the exploration which requires too much prerequisites and time.

To address these challenges, we propose a behavioral-based GUI testing approach for mobile applications that achieves faster and higher coverage. The proposed approach does not conduct static analysis. It creates a behavioral model from usage logs by applying a statistical model. The events within the behavioral model are mapped to GUI components in a GUI tree. Finally, it updates the model dynamically to increase the probability of an event that rarely or never occurs when users use the application. The proposed approach was evaluated on four open-source Android applications, and compared with the state-of-the-art tools and manual testing. The main evaluation criteria are code coverage and ability to find errors. The proposed approach performed better than the current state-of-the-art automated testing tools in most aspects.