Acronym:

Beyond the Nyquist limit 

From: 2015-10-01 To: 2019-09-30 (Completed)

HRZZ 

- 

prof. dr. sc. Damir Seršić

- 

Modern information and communication technologies face various challenges. We identified two of them as important: huge amount of data delivered from the real world via sensing devices, and computing power required for its processing. Both challenges are targeted in our project proposal, as well as some advanced applications of the proposed solutions. Recently, a very hot research topic named compressed sensing has been introduced. It is a technique to obtain relevant data from a significantly reduced set of measurements. Similar approach that deals with enormous amounts of data is sub-Nyquist sampling in a union of (analogue) subspaces. Both approaches result in applications that go beyond the known limits: they result in novel imaging techniques, advanced radar and sonar devices, wide-band software defined radios, as well as in advanced speech and audio processing systems. They rely on sparse representation of signals, sophisticated reconstruction and signal processing algorithms, as well as on intelligent hardware implementations that mix analogue and digital subsystems. The researchers gathered in this project team have experience in sparse signal modeling, in design of advanced analogue and digital systems, and have developed various applications in 3D, image, audio and speech signal processing. In this project, sparse representation based on novel parametric and non-parametric robustly adaptive systems will be used. Furthermore, inherently non-linear methods of modeling n-dimensional signals will be introduced. To increase the processing power, we will consider efficient software implementation of the algorithms on single-instruction-multiple-data platforms and their hardware implementation on field programmable gate arrays. The resulting methods and algorithms will contribute to rapidly increasing field of embedded systems. Moreover, new structures will be suitable for system-on-chip design. The results will be used in design of novel sub-Nyquist systems. Finally, we will develop its applications in 3D, image, audio and speech signal processing.