Compressed Sensing für Mobilfunknetze

• Applicant: Professor Dr. Rudolf Mathar
• Subject Area: Electronic Semiconductors, Components and Circuits, Integrated Systems, Sensor Technology, Theoretical Electrical Engineering
• Term: from 2011 to 2015
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 194709232

• Final Report Year: 2014

Final Report Abstract

The main goal of the project has been to exploit the benefits of compressed sensing for improving mobile communication systems. Within the broad field of mobile communications, two applications have been selected for investigation, namely relay networks and spectrum sensing. In relay networks, the goal has been to use compressed sensing as a distributed source coding algorithm and to optimize the relaying strategy such that the network compresses correlated transmitted signals, simply by forwarding them. We have shown that distributed compressed sensing can be performed reliably over multipleinput multiple-output (MIMO) multi-user access channels (MAC). A framework for analyzing the problem and a heuristic achieving near-optimal performance has been developed. In spectrum sensing, we intended to provide reliable spectrum occupation and channel state information from a drastically reduced number of measurements. In this area we have produced multiple results. We have developed an algorithm that exploits the fact that channel occupation changes relatively slowly. Using measurements from past sensing cycles, we employ matrix completion to recover the current frequency spectrum from a small number of measurements. Further, we have formulated a convex optimization program that both recovers as well as fuses sensing information from multiple sensors. Including the measurements from multiple sensors in a single recovery step makes it possible to drastically lower the number of measurements necessary for reliable frequency spectrum reconstruction. Exploiting the built-in periodicity inherently present in most man-made signals, we have formulated an optimization problem taking into account not only the knowledge about the received signal’s sparsity in the cyclic autocorrelation (CA) domain, but also the prior knowledge about the cyclic autocorrelation function’s (CAF) harmonic structure.

Publications

• “Distributed sensing of a slowly time-varying sparse spectrum using matrix completion,” in International Symposium on Wireless Communication 2011 (ISWCS 2011), Aachen, Deutschland, Nov. 2011, pp. 296–300
  S. Corroy, A. Bollig, and R. Mathar
  
• “Distributed compressed sensing for the MIMO MAC with correlated sources,” in IEEE International Conference on Communications (ICC 2012), Ottawa, Canada, Jun. 2012, pp. 2544–2548
  S. Corroy and R. Mathar
  
• “Joint sparse spectrum reconstruction and information fusion via 1-minimization,” in IEEE Vehicular Technology Conference 2012 Spring (VTC 2012- Spring), Yokohama, Japan, May 2012, pp. 1–5
  A. Bollig, S. Corroy, and R. Mathar
  
• “Dictionary-based reconstruction of the cyclic autocorrelation via l1-minimization,” in The 38th International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2013), Vancouver, Canada, May 2013, pp. 4908–4912
  A. Bollig and R. Mathar
  
• “MMME and DME: Two new eigenvalue-based detectors for spectrum sensing in cognitive radio,” in IEEE GlobalSIP 2013, Austin, Texas, U.S.A, Dec. 2013, pp. 1210–1213
  A. Bollig and R. Mathar