MAST

A reliable and accurate observation of the situation under water is essential to ensure the security of maritime infrastructures like harbors or sea cables. The previous research at the institute already looked at the technologies multiple input multiple output (MIMO) as well as artificial intelligence (AI) for sonar related applications. However, additional research fields need to be addressed to utilize the full potential of those technologies.

Goals and Methodology

Goal of the project MAST is the improvement of the capturing and processing of sonar data. For MIMO sonar antennas the main focus lies on the improvement of the signal separation through innovative pulse form design with a special attention on concrete applications like high resolution scans of ship hulls. In addition, when analyzing sonar images, not only the automatic detection of objects but also the detection of changes in a scene over a certain time period will be investigated in MAST. Here the focus lies on improving both analyses by means of modern deep learning methods.

Innovations and Perspectives

The project MAST fundamentally contributes to the development of MIMO and AO technologies for sonar systems and thus to the rise of the security of maritime infrastructures. Particularly relevant are research findings regarding the pulse form design since this is essential for the successful application of MIMO sonars. Furthermore, solutions regarding the training of deep learning models with a limited amount of data are also relevant outside the field of sonar image processing. In addition, the investigations highlight innovative application fields of the MIMO technology under water. The improved automatic analysis of sonar images leads besides others to a more effective search for unexploded ordnances or detection of changes on under water infrastructures.