Methodology
The achievements of ROVINA will be innovative solutions that allow for the digital preservation of unaccessible archeological sites with mobile robots. At the same time, we are extending the current state of the art in autonomous navigation, computer vision and intuitive human robot interfaces. The developments carried out in this project are the enabling technology to build autonomous robots that are able to explore challenging archeological sites including underground environments such as the catacombs of Rome. This requires us to address challenging problems in the context of 3D reconstruction and mapping under uncertainty, semantic analysis, object detection and online learning, environment analysis, autonomous navigation and exploration, as well as intuitive user interfaces. The best way to develop such a system with different partners is by means of a modular architecture, where each module fulfills a specific task and where the modules communicate with each other through a middleware with defined interfaces to exchange information. The modular structure simplifies the development of robust and reusable components. The modular architecture is exactly the approach that the consortium undertakes in ROVINA to bring this project to the desired success.
Every academic partner of the ROVINA is a world-leading expert in his field. The project is accompanied by ICOMOS-IT, the International Council on Monuments and Sites, with strong expertise in archeology and the digital preservation of such sites. The consortium also involves one SME, which is a highly skilled company
in the high-tech sector. Although having substantially different skills, there is enough overlap in the expertise of the consortium members to properly communicate with each other and to understand the other partners’ problems and challenges. The consortium members have a substantial record of joint collaboration. Moreover, the modular approach taken in ROVINA has been proven to be successful in the past and the integration efforts will start very early in the project, as soon as the first developments have begun. A proper integration of robust modules is important in robotics projects with a large degree of autonomous behavior, especially if the robots operate in inaccessible or hazardous environments as the ones addressed in this project. Only a seamlessly integrated system can sustain in these settings. Beside the integration activities that run concurrently with the developments, three integration weeks, where all partners meet to boost the integration of the individual components, will be carried out during the project at months 15, 24 and 36. The results of these efforts are then directly visible in the subsequent evaluations of the ROVINA system.
An element of ROVINA is the fact that all key developments will be published as open source software as well as under a commercial license. Inspired by companies such as Trolltech or Nokia, that release software such as the Qt framework under an open source and a commercial license, ROVINA will apply a dual licensing model. First, using a standard open source licensing scheme such as GPL, L-GPL or BSD and second, providing commercial licenses that typically better fit the needs of the industry. As a result, all developed components will be made available to the public and will in this way disseminate the results of ROVINA in a much broader way compared to public deliverables only as used in most European projects. Furthermore, the open source release will allow other groups to reduce the cost for building robot-based digital preservation products in the future. The ROVINA consortium believes that this decision will substantially increase the impact of the project and its results and will also trigger a series of other, new robotics applications, since modern, state-of-the-art components can directly be used by others. It regards this comparably open strategy as appropriate, since large parts of the funding comes in the end from European tax payers. The software developments will be made available in a ROVINA software repository accessible via the internet.