SUMMARY

This proposal concerns the study and development of methodologies for optimising acoustical and optical sensors' functioning and integrating related data for the formation of an accurate virtual environment aimed at supporting navigation, inspection, and maintenance/repair tasks of a multifunctional remotely operated underwater vehicle (ROV).

At present, the guidance and the inspection tasks performed by ROVs are not easy for several reasons.
For instance, every kind of data acquired by a sensor device on-board a vehicle is controlled by an operator.
The ROV navigation is made difficult by the lack of depth perception of the operator who utilises, typically, only optical cameras. In other words, the several sensors work separately and for different goals. As a consequence, to perform an effective operation, an ROV requires specialised crew, expensive training course and many hours of practice.

The integrated use of acoustical and optical cameras to reconstruct the actual scenario by a virtual environment improves the pilot comfort and then speeds up the operations, also reducing the number of operators and then the overall cost.
Moreover, this sharply increases the capabilities of telerobotic tasks in that an operator can interact with the synthetic scenario, understanding the sur-rounding situation, navigating and simulating a specific task before actually per-forming it. Also, deskilling the operator by the use of better computer model based controller will reduce the cost of ROV operations.

In this context, this proposal aims at achieving mainly two specific targets:

• to prove the feasibility of a new high-performance sensing and processing system for underwater imaging, based on optical and acoustical sensors and in their intelligent and integrated use to provide precise two- and three-dimensional information;

• the development of a 3D virtual environment from sensorial data reproducing the observed scene, supporting an operator to monitor the actual situation and control the vehicle in all environmental conditions.

These objectives can be attained by:

• using multisensorial imaging devices, i.e., optical and acoustical cameras, providing real-time data and appropriate methodologies (e.g., sensor integration, data fusion, computer vision approaches) to process and interpret such data;

• using novel methods for building and updating the virtual environment at different resolution;

• developing adequate modelling and simulation procedures to emulate typical and non-typical operations.

• Technical Software Consultant  (TSC) Ltd, United Kingdom (coordinator)

• OMNITECH (OT) A/S, Norway

• University of Genova, Dipartimento di Scienze dell'Informazione (DISI), Italy

Link to the online demo   (Cosmo Player  plug-in required!)

Official VENICE web page