Compressive Sensing Multiuser-Detection for Code-Multiplex-Systems

• Applicant: Professor Dr.-Ing. Armin Dekorsy
• Subject Area: Electronic Semiconductors, Components and Circuits, Integrated Systems, Sensor Technology, Theoretical Electrical Engineering
• Term: from 2011 to 2017
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 204084647

For a summary of the general objectives and characteristics of this project, please be referred to the original project proposal, pp. 5.During the granted project period of two years considerable parts of the work planned for a three year period have been finished. Specifically, AP1 has been fully treated and AP2 has been intensively treated applying Greedy Compressive Sensing (CS) algorithms in deviation of the original work plan. This change of focus in the applied CS algorithms has been mainly motivated by a lower complexity and higher adaptability of known Greedy algorithms for the purpose of data detection. A detailed discussion of this issue can be found in the intermediate report. The original AP3 has not been work on due to the granted project duration of two years only. However, the already achieved results of the project indicate that the original AP3 focus has to be adapted towards soft decisions for data as well as activity detection. Only by also processing activity reliability information effective iterative detection schemes will be enabled. As a consequence, AP3 will be replaced by a new working package according to the project observations and newly developed research questions. Finally, AP4 has been generally treated, but will be extended during the second project phase to provide a systematic comparison of developed algorithms and results for system dimensioning. In addition to the original work plan a new AP5 has been added to treat channel estimation. Previous results and recent theoretic insights from CS literature indicate that a joint estimation of channel, data and activity is principally possible. Therefore, the focus of AP5 will be on development of suitable estimation algorithms.During the two years of the project five publications, of which one is a journal article, have been published. Furthermore, three student theses have been finished and the topic Compressed Sensing has been included in lectures.

• DFG Programme: Research Grants
• Participating Person: Dr.-Ing. Carsten Bockelmann