Variability Modeling and Resolution in Component-based Robotics Systems

The routine use of existing solutions in the development of new systems is a key attribute of every mature engineering discipline. Software reuse is indeed the state of the practice in various domains (e.g. telecommunications, automotive and avionics). One of the key factors that enable the development of reusable software is the flexibility, which is the ease with which a system can be modified to be used in applications or environments different from those for which it was originally designed. In robotics software reuse is still at an early stage. This is due to the complexity and the huge variability that characterize this domain. The complexity makes the development of reusable software a task that requires advanced software engineering techniques, which are not always mastered by robotics experts. The variability of hardware, environment and task makes frequent the changes that occur in the application requirements. The research documented in this thesis investigates new approaches for the development of component-based robotics systems, which are flexible enough to accommodate the changes that are likely to occur to the software requirements. Addressing the flexibility of robotics software systems and variability modeling and resolution are thus the main topics. The first contribution is a software development process that explicitly takes into account the variability. The process is based on two of the most promising approaches to software reuse, namely the Software Product Lines and the Model Driven Engineering. This thesis describes the process, illustrates the models and tools that have been developed for supporting it, and exemplifies its application by means of a case study. The second contribution consists of a set of approaches that have been designed for developing component-based system with an high level of flexibility and reusability. In particular this approaches focus on addressing the software framework and the hardware variability.