About this Journal Submit a Manuscript Table of Contents
International Journal of Antennas and Propagation
Volume 2012 (2012), Article ID 713594, 8 pages
http://dx.doi.org/10.1155/2012/713594
Research Article

A New Generation Method for Spatial-Temporal Correlated MIMO Nakagami Fading Channel

1College of Electronic Information Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China
2School of Software, Tsinghua University, Beijing 100084, China
3National Mobile Communications Research Laboratory, Southeast University, Nanjing 210008, China

Received 10 July 2011; Accepted 5 September 2011

Academic Editor: Yan Zhang

Copyright © 2012 Qiu-Ming Zhu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

A new generation method for spatial and temporal correlated multiple-input multiple-output (MIMO) Nakagami fading channel is proposed, which has low complexity and is applicable for arbitrary fading parameters and prespecified correlation coefficients of different subchannel. The new scheme can be divided into two steps: (1) generate independent Nakagami fading sequences for each subchannel based on a novel rejection method; (2) introduce the temporal and spatial correlation based on the relationships between Rayleigh, Gamma, and Nakagami random processes. The analysis and simulation results show that the proposed simulator has a good agreement with the theoretical model on fading envelope distribution, spatial-temporal correlation characteristic.

1. Introduction

In the past decades, more and more attentions have been given to MIMO system for its high channel capacity [1]. Many studies and measured data also showed channel fadings on different antennas are correlated. Moreover, the Nakagami fading encompasses Rayleigh and Rice fadings and shows good performance in matching experimental data with different fading parameter [2]. Hence, correlated MIMO channel modeling and capacity evaluation become the interest of many researchers recently [3, 4]. However, few published works focus on simple and accurate generating of spatial-temporal correlated MIMO Nakagami fading channel which is important for design or laboratory test of MIMO systems.

A number of methods for generating independent Nakagami fading channels are available in the literature. Yip and Ng’s method [5] based on the product of square-root beta random variates (RVs) and Rayleigh RVs is accurate but only valid for . The decomposition method [6] is also only valid for integer or half-integer value of . Beaulieu and Cheng’s method [7] based on the inverse cumulative distribution function of Nakagami distribution is suitable for arbitrary but requires a different set of coefficients by numerical computation or curve fitting. The rejection method in [8, 9] is accurate and universal but needs a set of constant coefficients determined empirically or lacks efficiency. In light of this fact, a novel simple and high-efficient generation method for arbitrary is proposed in our simulator which yields an excellent accuracy.

Yip and Ng’s method and the decomposition method are also suitable for generating temporal correlated Nakagami fadings with specific values of . Recently, Ma and Zhang [10] proposed a universal method by constructing a new map relation between Gaussian and Nakagami sequences, but it is quite difficult to calculate the autocorrelation of the corresponding input Gaussian sequence. In [11], a simple method called rank-matching was proposed to introduce prespecified temporal correlation by rearranging the independent Nakagami sequence, while the author addressed the relationship between Rayleigh and Nakagami envelope in an approximation manner. In [6, 12, 13], techniques of channel correlation matrix decomposition for generating spatial correlated channels were reported.

In this paper, we propose a new framework for generating the correlated MIMO Nakagami fading channels which possess (1) arbitrary and different for each subchannel; (2) arbitrary and different average fading power; (3) arbitrarily prespecified temporal correlation coefficients; (4) arbitrary spatial correlation coefficient matrix. The new method consists of two steps: generate independent Nakagami sequence and introduce spatial-temporal correlation. Additionally, the simulation error is also analyzed and testified by numerical examples.

The rest of the paper is organized as follows. In Section 2, a general description of MIMO channel is given. In Section 3, the new framework for simulating the correlated Nakagami fading MIMO channel is described and analyzed. In Section 4, the generation approach is verified through numerical simulation. Finally, the conclusions are drawn in Section 5.

Throughout the paper, we use to denote transposition. and means the expectation and variance, respectively. is the temporal correlation coefficient (TCC) of the th subchannel, and is the spatial correlation coefficient (SCC) between subchannel and . denotes the Gamma distribution with distribution parameters and .

2. MIMO Channel Model

2.1. General Description

A flat fading MIMO system equipped with transmit and receive antennas is showed in Figure 1. The transmit and receive signals are expressed by vectorized forms, and . The input-output relationship is given by where corresponds to the additive noise matrix, is the time-variate channel fading matrix which can be expressed as where is the complex fading coefficient from th transmit antenna to th receive antenna. For simplicity, is transformed to a column vector containing elements, or The envelope of follows Nakagami distribution, and the phase is uniformly distributed in . The TCC of each subchannel is denoted by , and the SCC matrix of can be described by a matrix .

713594.fig.001
Figure 1: MIMO system with transmit and receive antennas.
2.2. Nakagami, Rayleigh, and Gamma Distribution

The probability density function (PDF) of Nakagami distribution was given in [2] as where is the mean power, denotes the gamma function, and the fading parameter controls the fading degree of the propagation field.

Following [2], the squared Nakagami RV obeys a Gamma distribution and can be constructed from the sum of independent and identically distributed (i.i.d) squared Rayleigh RVs, where and denote Rayleigh and Gamma RV, respectively, whose PDFs are defined as where .

3. Generation of Correlated Nakagami MIMO Channel

In general, the correlated Nakagami MIMO channel simulation is to generate Nakagami fading channels which are temporal and spatial correlated. It is assumed that the fading parameters and correlation properties are arbitrary and different.

3.1. Generating Independent Nakagami Fading

The rejection/acceptance technique is a universal generation method for nonuniform RVs, which is summarized as follows [14](1)Find a hat function which satisfies the inequality where is the required distribution and is the rejection constant.(2)Generate a RV with PDF and a RV uniformly distributed in .(3)If , is accepted. Otherwise, it is rejected.

The main difficulty of this technique lies in finding a hat function which is close to the required distribution and satisfies the inequality. In this paper, we adopt a new truncated Gaussian distribution function as the hat function for arbitrary . Using (8) and strict mathematical derivation, the hat function and rejection constant are founded as where , , and

An important measure for the quality of the rejection algorithm is the acceptance probability or rejection efficiency which involves the closeness between the Nakagami PDF and the hat function. Figure 2 plots the rejection efficiency curve of our proposed method. It shows that the efficiency decreases for high values of but better than 70%. However, the rejection efficiency of [8] is only 65.75% and 66.67% in cases of and , while the method in [9] becomes inefficient as increases and the efficiency is below 64% mostly.

713594.fig.002
Figure 2: Rejection efficiency of the new method.
3.2. Introducing Temporal Correlation

For stationary random processes, the TCC of the samples is well approximate with the TCC of their statistical ranks [15]. Reference [11, Equation (6)] gives the assumption without proving that the TCC of Nakagami envelope approximately equal to the corresponding Rayleigh envelope which means the TCCs of their statistical ranks are almost same.

The Nakagami envelope autocorrelation function is given by [2]: where is the autocorrelation coefficient of squared Rayleigh envelope and is the hypergeometric function. The statistical properties of Nakagami sequence can be found as

According to the definition of autocorrelation coefficient, we obtain Following [16], we have where is the autocorrelation coefficient of Rayleigh envelope.

It is difficult to get the accurate relationship expression between and . A numerical method is used to investigate the absolute error for different . The numerical results are plotted in Figure 3, which shows the maximum difference between two TCC curves is less than 0.025 which is fairly small for most practical purposes.

713594.fig.003
Figure 3: The absolute error between and .

Thus, we can rearrange the Nakagami sequence to match the statistical rank of the Rayleigh sequence in order to introduce the desired temporal correlation (see Figure 4). Moreover, there are many efficient methods to generate the reference Rayleigh sequence (or complex Gaussian sequence) with desired autocorrelation, such as the AR model [17] and the MEPA method [18]. It should be emphasized that the temporal correlated complex Gaussian sequence here can be transformed to the truncated Gaussian sequence for rejecting in (9).

713594.fig.004
Figure 4: Introducing temporal correlation.
3.3. Introducing Spatial Correlation

We introduce the spatial correlation for Nakagami fading subchannels following the flowchart in Figure 5. The key step of this procedure is generating correlated Gamma RVs whose square root yields the desired correlated Nakagami MIMO channel. There are two commonly used methods, the decomposition method [12] and the Simon method [19]. Although the Simon method shows better accuracy, it is more complex and has some constraints on the prespecified statistical parameters. The decomposition method is more versatile and simpler, which can be expressed as where is the normalized Gamma RV vector, and , is the desired cross-correlated Gamma RV vector, and denotes a low triangular matrix, which is the Cholesky decomposition on the covariance matrix of , It can be derived that , , are determined by the specified simulation parameters,

713594.fig.005
Figure 5: Introducing spatial correlation.

The decomposition method assumes that the linear summation of independent Gamma RVs follows Gamma distribution. In the following, we will analyze the simulation error caused by this assumption. Firstly, the th cross-correlated Gamma RV of (15) can be written as where denotes a new independent Gamma RV which follows , and the new distribution parameters yield

The exact PDF of can be found as [20, Equation (2.9)]: where ,, , and where .

The straightforward transformation leads the simulation PDF of Nakagami RV as

Finally, the theoretical expression for absolute error of simulated Nakagami PDF is obtained as Figure 6 compares the simulated results of Nakagami PDF error with the values calculated from the exact expression (23). In the simulation, 106 samples of the Nakagami RV are generated by four independent Gamma RVs with different parameters sets of and . It can be seen that a very good agreement between the simulated results and the exact expression is obtained. Anyway, the absolute error of Nakagami PDF is very small which means the decomposition method can achieve satisfactory performance.

713594.fig.006
Figure 6: Comparison of the absolute error of Nakagami PDF.

4. Simulation Results

To illustrate the behavior of the proposed channel simulator, we consider a MIMO system with at the mobile station (), at the base station (). The fading characteristics of four subchannels are different:

The is assumed to be surrounded by several scatterers which results in a low correlation between two antennas while the antennas of set high and have no local scatters, Hence, the SCC matrix is given by [21]: where denotes Kronecker product.

A plausible model for the angle of arrival (AOA) of the receiving signals is Von Mise distribution, which is defined as [22]: Under this scatting environment, the theoretical expression of TCC is given by [22] as where is the zero-order modification Bessel function, denotes the mean direction of AOA, controls the angular spread, and is the maximum Doppler frequency. Antenna arrays at both ends are assumed to be small spacing (less than half wavelength), so the temporal correlation properties of receiving signals on each receiving antenna are almost same.

In the simulation, 106 samples of each subchannel are generated based on the proposed scheme. Figure 7 shows that the simulated PDFs of fading envelope match the theoretical curves very well under four fading scenarios. The corresponding statistical parameters are estimated by maximum likelihood method as which is also very close to the desired ones.

fig7
Figure 7: The PDFs of simulated subchannel fading envelope.

We next look at the spatial and temporal correlation coefficients of the above outputs. In the simulation, we assume that four subchannels experience almost the same no-isotropic ( ) scattering scenarios for small antenna spacing. The TCC of each output channel is showed in Figure 8, where means the normalized time delay. It is showed that the simulation curves agree well with the desired ones. The SCC matrix is calculated by output samples ending up with Comparing to (27), it is found that the relative errors are 3.54%, 2.47%, and 1.27%, respectively.

fig8
Figure 8: The TCCs of simulated subchannel fading envelope.

5. Conclusions

In this paper, a new framework for simulating the spatial and temporal correlated MIMO Nakagami fading channels was proposed. It can be applied on arbitrary fading parameters and prespecified spatial-temporal correlation coefficients with a satisfactory accuracy. Moreover, the new simulation method is low complexity and applicable for large-scale and real-time channel simulation scenarios. The most complex part of this simulator is to generate complex Gaussian sequences which can be accomplished by well-established methods.

Acknowledgments

This work is supported by National Nature Science Foundation of China (no. 61102068), the Aviation Science Foundation of China (no. 2009ZC52036), Nanjing University of Aeronautics and Astronautics Research Funding (no. NP2011036 and NS2011013), and China Postdoctoral Science Foundation funded project (no. 20110490389).

References

  1. E. Telatar, “Capacity of multi-antenna Gaussian channels,” European Transactions on Telecommunications, vol. 10, no. 6, pp. 585–595, 1999. View at Scopus
  2. M. Nakagami, The m-Distribution: A General Formula of Intensity Distribution of Rapid Fading Statistical Methods of Radio Wave Propagation, W.C. Hoffman, Japan, 1960.
  3. M. Matthaiou, D. I. Laurenson, and J. S. Thompson, “A MIMO channel model based on the Nakagami-faded spatial eigenmodes,” IEEE Transactions on Antennas and Propagation, vol. 56, no. 5, pp. 1494–1497, 2008. View at Publisher · View at Google Scholar · View at Scopus
  4. C. Zhong, K. K. Wong, and S. Jin, “Capacity bounds for MIMO Nakagami-m fading channels,” IEEE Transactions on Signal Processing, vol. 57, no. 9, pp. 3613–3623, 2009. View at Publisher · View at Google Scholar · View at Scopus
  5. K. W. Yip and T. S. Ng, “A simulation model for nakagami-m fading channels, m < 1,” IEEE Transactions on Communications, vol. 48, no. 2, pp. 214–221, 2000. View at Scopus
  6. Q. T. Zhang, “Decomposition technique for efficient generation of correlated Nakagami fading channels,” IEEE Journal on Selected Areas in Communications, vol. 18, no. 11, pp. 2385–2392, 2000. View at Publisher · View at Google Scholar · View at Scopus
  7. N. C. Beaulieu and C. Cheng, “Efficient nakagami- fading channel simulation,” IEEE Transactions on Vehicular Technology, vol. 54, no. 2, pp. 413–424, 2005. View at Publisher · View at Google Scholar · View at Scopus
  8. L. Cao and N. C. Beaulieu, “Simple efficient methods for generating independent and bivariate Nakagami-m fading envelope samples,” IEEE Transactions on Vehicular Technology, vol. 56, no. 4 I, pp. 1573–1579, 2007. View at Publisher · View at Google Scholar · View at Scopus
  9. M. Matthaiou and D. I. Laurenson, “Rejection method for generating Nakagami-m independent deviates,” Electronics Letters, vol. 43, no. 25, pp. 1474–1475, 2007. View at Publisher · View at Google Scholar · View at Scopus
  10. Y. Ma and D. Zhang, “A method for simulating complex nakagami fading time series with nonuniform phase and prescribed autocorrelation characteristics,” IEEE Transactions on Vehicular Technology, vol. 59, no. 1, Article ID 5223623, pp. 29–35, 2010. View at Publisher · View at Google Scholar · View at Scopus
  11. J. C. S. Silveira Santos Filho and M. D. Yacoub, “On the simulation and correlation properties of phase-envelope nakagami fading processes,” IEEE Transactions on Communications, vol. 57, no. 4, pp. 906–909, 2009. View at Publisher · View at Google Scholar · View at Scopus
  12. K. Zhang, Z. Song, and Y. L. Guan, “Simulation of nakagami fading channels with arbitrary cross-correlation and fading parameters,” IEEE Transactions on Wireless Communications, vol. 3, no. 5, pp. 1463–1468, 2004. View at Publisher · View at Google Scholar · View at Scopus
  13. L. C. Tran, T. A. Wysocki, and J. Seberry, “A generalized algorithm for the generation of correlated Rayleigh fading envelopes in radio channels,” in Proceedings of the 19th IEEE International Parallel and Distributed Processing Symposium, (IPDPS'05), vol. 13, p. 238, Denver, Colo, USA, 2005.
  14. A. Papoulis and S. U. Pillai, Probability Random Variables and Stochastic Processes, McGraw-Hill, Boston, Mass, USA, 4th edition, 2002.
  15. E. L. Lehmann and H. L. M. D'Abrera, Nonparametrics: Statistical Methods Based on Ranks, McGraw-Hill, New York, NY, USA, 1975.
  16. W. C. Jakes, Microwave Mobile Communications, Wiley, New York, NY, USA, 1974.
  17. K. E. Baddour and N. C. Beaulieu, “Autoregressive modeling for fading channel simulation,” IEEE Transactions on Wireless Communications, vol. 4, no. 4, pp. 1650–1662, 2005. View at Publisher · View at Google Scholar · View at Scopus
  18. Q. M. Zhu, D. Z. Xu, and X. M. Chen, “Modified channel model for non-isotropic scattering environments,” in Proceedings of the International Conference on Wireless Communications & Signal Processing, (WCSP '09), pp. 1–5, Nanjing, China, November 2009.
  19. C. H. Simon, “Generation of Poisson and Gamma random vectors with given marginal and covariance matrix,” Journal of Statistical Computation and Simulation, vol. 47, no. 1, pp. 1–10, 1993.
  20. P. G. Moschopoulos, “The distribution of the sum of independent gamma random variables,” Annals of the Institute of Statistical Mathematics, vol. 37, no. 1, pp. 541–544, 1985. View at Publisher · View at Google Scholar · View at Scopus
  21. J. P. Kermoal, L. Schumacher, K. I. Pedersen, P. E. Mogensen, and F. Frederiksen, “A stochastic MIMO radio channel model with experimental validation,” IEEE Journal on Selected Areas in Communications, vol. 20, no. 6, pp. 1211–1226, 2002. View at Publisher · View at Google Scholar · View at Scopus
  22. A. Abdi, J. A. Barger, and M. Kaveh, “A parametric model for the distribution of the angle of arrival and the associated correlation function and power spectrum at the mobile station,” IEEE Transactions on Vehicular Technology, vol. 51, no. 3, pp. 425–434, 2002. View at Publisher · View at Google Scholar · View at Scopus