Loading...

Joint Pilot Power & Pattern Design for Compressive OFDM Channel Estimation

Khosravi, Mahdi | 2014

2085 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 46013 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Mashhadi, Saeed
  7. Abstract:
  8. Increasing need to high rates of transmission through radio channels is a challenging problem in wireless communications. Simple implementation and high spectral efficiency of OFDM system turned it into a suitable choice to meet this need. Accurate estimation of the communication channel has a significant impact on performance of Orthogonal Frequency Division Multiplexing (OFDM) systems. Conventional methods of channel estimation are not able to exploit the inherent sparsity of the channel that is due to sparse distribution of scaterrers. On the other hand, Compressed Sensing (CS) is getting attention in variety of aspects such as communications, recently.CS-based channel estimation exploits sparsity of the channel to reduce the number of pilots and increase the spectral efficiency. In this thesis, channel estimation techniques in OFDM systems are reviewed. Then, a basic introduction to CS is given. Afterwards, CS-based channel estimation techniques are categorized and introduced.Next, we tried to propose a method to obtain an optimal pattern and power of pilots in OFDM systems based on minimizing the coherency of the sensing matrix. The sensing matrix in this problem is product of a diagonal matrix whose elements are power of pilots by a Discrete Fourier Transform (DFT) sub-matrix whose rows are selected according to pilot pattern. The proposed algorithm is restricted to pilot patterns forming Cyclic Difference Sets or Almost Difference Sets. Once a pattern is found it uses Nelder-Mead numerical optimization method to calculate the optimal power of pilots. Afterwards, the search is performed again to find more patterns. Finding each pattern is followed by calculating the power allocation among its pilots.Finally, the pair leading to minimum coherence is selected. Meanwhile, the algorithm is provided with a parameter specified by user to limits its running time. Simulations
    show that the proposed scheme has 3 dB efficiency over the existing schemes. Also, its convergence speed is higher than that of the counterparts
  9. Keywords:
  10. Channel Estimation ; Compressive Sensing ; Optical Orthogonal Frequency Division Multiplexing (OFDM) ; Pilot Design ; Cyclic Difference Set (CDS) ; Almost Difference Set (ADS)

 Digital Object List

 Bookmark

...see more