No abstract

The development of software for Wireless Sensor Networks (WSN) is mostly based on code-and-fix techniques. Up today model-driven engineering techniques have only been limitedly considered in this domain, although they would enable a set of activities aimed at improving the quality of software. In this paper we investigate the possibility to adapt an existing model-based approach that exploits such techniques to combine the modeling and performance analysis of software for WSN. We introduce a UML-based framework where a system model (i) is extended with a new profile for representing NesC application along with the supporting hardware platform, and (ii) is annotated with performance parameters defined in the standard UML MARTE profile. Thereafter we apply a set of transformations to this enhanced UML model that targets a Queueing Network performance model. Finally we illustrate our approach at work on a case study in the agricultural domain.

Abstract. Wireless Sensor Networks are becoming one of the most successful choices for the development and deployment of a wide range of applications, from intelligent homes to environment monitoring. In response to the growing demand for fast development of WSN applications, we extend an existing UML-based approach for the design and code generation of Agilla applications with functional simulation and timing analysis capabilities through executable UML models. The proposed approach makes use of both a UML profile and an executable model library for Agilla. Execution times, annotated on Agilla instructions and patterns in the library, are given as additional input parameters during the model execution to carry out a timing analysis of the simulated Agilla applications. Modeling and simulation activities rely on MagicDraw© and Cameo Simulation Toolkit©. A running case study is provided to show the approach and the supporting tools at work.

No abstract

Wireless Sensor Networks (WSN) are nowadays applied to a wide set of domains (e.g., security, health). WSN are networks of spatially distributed, radio-communicating, battery-powered, autonomous sensor nodes. WSN are characterized by scarcity of resources, hence an application running on them should carefully manage its resources. The most critical resource in WSN is the nodes' battery.In this paper, we propose model-based engineering facilities to analyze the energy consumption and to develop energy-aware applications for WSN that are based on Agilla Middleware. For this aim i) we extend the Agilla Instruction Set with the new battery instruction able to retrieve the battery Voltage of a WSN node at run-time; ii) we measure the energy that the execution of each Agilla instruction consumes on a target platform; and iii) we extend the Agilla Modeling Framework with a new analysis that, leveraging the conducted energy consumption measurements, predicts the energy required by the Agilla agents running on the WSN. Such analysis, implemented in fUML, is based on simulation and it guides the design of WSN applications that guarantee low energy consumption. The approach is showed on the Reader agent used in the WildFire Tracker Application.