Asad MAHMOOD, Abdul GHAFOOR, Sajjad HUSSAIN, Nadia IQBAL

Bayesian compressive sensing framework for spectrum reconstruction in Rayleigh fading channels

Compressive sensing (CS) is a novel digital signal processing technique that has found great interest in many applications including communication theory and wireless communications. In wireless communications, CS is particularly suitable for its application in the area of spectrum sensing for cognitive radios, where the complete spectrum under observation, with many spectral holes, can be modeled as a sparse wide-band signal in the frequency domain. Considering the initial works performed to exploit the benefits of Bayesian CS in spectrum sensing, the fading characteristic of wireless communications has not been considered yet to a great extent, although it is an inherent feature for all sorts of wireless communications and it must be considered for the design of any practically viable wireless system. In this paper, we extend the Bayesian CS framework for the recovery of a sparse signal, whose nonzero coefficients follow a Rayleigh distribution. It is then demonstrated via simulations that mean square error significantly improves when appropriate prior distribution is used for the faded signal coefficients and thus, in turns, the spectrum reconstruction improves. Different parameters of the system model, e.g., sparsity level and number of measurements, are then varied to show the consistency of the results for different cases.

PDF

___

[1] Mitola J, Maguire GQ. Cognitive radio: making software radios more personal. IEEE Pers Commun 1999; 6: 13-18.
[2] Shankar NS, Cordeiro C, Challapali K. Spectrum agile radios: utilization and sensing architectures. In: 2005 First IEEE International Symposium on New Frontiers in Dynamic Spectrum Access Networks; 811 November 2005; Baltimore, MD, USA. pp. 160-169.
[3] Tang H. Some physical layer issues of wideband cognitive radio systems. In: 2005 First IEEE International Symposium on New Frontiers in Dynamic Spectrum Access Networks; 811 November 2005; Baltimore, MD, USA. pp. 151-159.
[4] Cabric D, Mishra SM, Brodersen RW. Implementation issues in spectrum sensing for cognitive radios. In: Conference Record of the Thirty-Eighth Asilomar Conference on on Signals, Systems and Computers; 710 November 2004; Pacific Grove, CA, USA. pp. 772-776.
[5] Candes EJ, Romberg J, Tao T. Robust uncertainty principles: exact signal reconstruction from highly incomplete frequency information. IEEE T Inform Theory 2006; 52: 489-509.
[6] Donoho D. Compressed sensing. IEEE T Inform Theory 2006; 52: 1289-1306.
[7] Sarvotham S, Baron D, Baraniuk R. Compressed Sensing Reconstruction Via Belief Propagation. Technical Report. Houston, TX, USA: Rice University, 2006.
[8] Baron D, Sarvotham S, Baraniuk R. Bayesian compressive sensing via belief propagation. IEEE T Signal Proces 2010; 58: 269-280.
[9] Gallager RG. Low-density parity check codes. IRE T Inform Theor 1962; 8: 21-28.
[10] MacKay DJC. Good error-correcting codes based on very sparse matrices. In: IEEE International Symposium on Information Theory; 29 June-4 July 1997; Ulm, Germany. p. 113
[11] Tan X, Li J. Computationally efficient sparse Bayesian learning via belief propagation. IEEE T Signal Proces 2010; 58: 2010-2021.
[12] Akcakaya M, Park J, Tarokh V. A coding theory approach to noisy compressive sensing using low density frame. IEEE T Signal Proces 2011; 59: 5369-5379.
[13] Kang J, Lee HN, Kim K. On detection-directed estimation approach for noisy compressive sensing. 2012 [Online]. arXiv: 121 1.1250 [cs.IT].
[14] Kang J, Lee HN, Kim K. Detection directed sparse estimation using bayesian hypothesis test and belief propagation. 2012 [Online]. arXiv: 121 1.1250 [cs.IT].
[15] Ishwaran H, Rao JS. Spike and slab variable selection: frequentist and Bayesian strategies. Ann Stat 2005; 33: 730-773.
[16] Ji Y, Yang Z, Li W. Bayesian sparse reconstruction method of compressed sensing in the presence of impulsive noise. Circ Syst Signal Pr 2013; 32: 2971-2998.
[17] Wang YG, Yang L, Tang L, Liu Z, Jiang WL. Enhanced multi-task compressive sensing using Laplace priors and MDL-based task classification. EURASIP J Adv Sig Pr 2013; 160: 1-17.
[18] He L, Carin L. Exploiting structure in wavelet-based Bayesian compressive sensing. IEEE T Signal Proces 2009; 57: 3488-3497.
[19] Guo D, Wang CC. Asymptotic mean-square optimality of belief propagation for sparse linear system. In: IEEE Information Theory Workshop; 2226 October 2006; Chengdu, China: pp. 194-198.
[20] Wainwright MJ, Jaakkola T, Willsky A. Tree-based reparameterization framework for analysis of sum-product and related algorithms. IEEE T Inform Theory 2003; 49: 1120-1146.
[21] Rangan S. Estimation with random linear mixing, belief propagation and compressed sensing. In: 44th Annual Conference on Information Sciences and Systems; 1719 March 2010; Princeton, NJ, USA. pp. 1-6.