Motivation:

At our department we have developed the approach 2D Compressed Edge Spectrum Sensing for Cognitive Radio applications. Here, we make use of the sparsity of spectrum edges to sense a wideband of interest at a considerable lower cost. The description of the corresponding problem differs somehow compared to the classical Compressed Sensing approach. Therefore, we need new algorithms providing low computational complexity and latency.

Goal:

Your task is to implement several analysis cosparsity reconstruction algorithms of the research literature and to compare them regarding complexity and reconstruction time. The challenge of this thesis is to make them efficient as possible. If good ideas come into your mind, you are free to create your own algorithm.

Requirements:

Deep understanding of stochastics and linear algebra is required. Successful participation in digital signal processing is advantageous.