 Open Access
 Total Downloads : 109
 Authors : Sameer Kumar, Shikha Singh
 Paper ID : IJERTV3IS061632
 Volume & Issue : Volume 03, Issue 06 (June 2014)
 Published (First Online): 27062014
 ISSN (Online) : 22780181
 Publisher Name : IJERT
 License: This work is licensed under a Creative Commons Attribution 4.0 International License
Carrier Frequency Synchronization in MIMO Systems
Sameer Kumar
M.Tech. Student, Department of Electronics And Telecommunication Engineering Dr.C.V.Raman University ,(CVRU)
Kota ,Bilaspur (C.G.), India
Shikha Singh
Asst. Professor
Electronics and Telecommunication Engineering, Dr.C.V.Raman University,(CVRU)
Kota, Bilaspur (C.G.), India
AbstractThis paper gives a summary of the principles, key techniques and state of the art of frequency synchronization in multiinput multioutput (MIMO) systems. The system model of frequency synchronization in MIMO systems is given and some assumptions are made first, and the principles and state of the art of the estimation and compensation of carrier frequency offsets (CFOs) are discussed for colocated MIMO and distributed MIMO systems, respectively. By analyzing the disadvantages of the existing research, we present some open issues worth of studying and some promising research interest. The results of analysis show that there has been adequate intensive study on the CFO estimation for colocated MIMO systems while the CFO estimation for distributed MIMO systems and the CFO compensation for MIMOOFDM (orthogonal frequency division multiplexing) systems appear incomplete and still need further research..
KeywordsCarrier frequency offsets; Distributed antennas; Frequency synchronization; MIMO; OFDM.

INTRODUCTION
In carrier communication systems, it is the task of the local oscillator in the receiver to generate a local reference carrier, whose frequency must be exactly equal to that of the carrier generated by the transmitter, to demodulate the received signal. Unfortunately, in practical communication systems, there always exist carrier frequency offsets (CFOs) between the transmitter and the receiver, which is unavoidably present due to the possible oscillator mismatch as well as the Doppler shifts caused by the relative motion between the transmitter and the receiver. The presence of CFOs will lead to significant performance degradation in the signal detection and the estimation of other important signal parameters (e.g., channel estimation) as well [1]. Therefore, it is of primary importance to accurately estimate the CFOs and compensate for them prior to detection, which is the main goal of frequency synchronization. The technology of frequency synchronization for conventional singleantenna systems has been studied intensively and adopted widely and will not be discussed here.
As compared with singleantenna systems, multi inputmultioutput (MIMO) systems can provide significantly higher data rate without extra bandwidth [2]. On the other
hand, orthogonal frequency division multiplexing (OFDM) can effectively combat the multipath fading of the wireless channel [3]. Therefore, the combination of MIMO and OFDM, resulting in a MIMOOFDM system, can provide higher capacity and achieve better performance [4]. As another promising technique for future wireless communications, systems with distributed antennas have been attracting much research interest in recent years [5].
Like that of singleantenna systems, the performance of MIMO systems may seriously degrade in the presence of CFOs. In MIMO systems, two possible approaches may be taken to deal with the CFOs. One is based on the elaborate design of the transmitted signal at the transmitter, where some encoding scheme immune to CFOs is used to combat the CFOs and hence the signal detection can be performed as if there were not any CFOs at the receiver [6]. The alternative approach is similar to that adopted in singleantenna systems, that is, frequency synchronization at the receiver, which can be carried out in two steps, i.e., CFO estimation and CFO compensation. The CFOs is first estimated by some CFO estimation method and then compensated for using the estimated CFO values obtained in the first step. We will focus on the second approach, i.e., frequency synchronization in MIMO systems.
This paper gives a summary of the principles, key techniques and state of the art of frequency synchronization in MIMO systems. The system model of frequency synchronization in MIMO systems is given and some assumptions are made first, and the principles and state of the art of CFO estimation and CFO compensation are discussed for colocated MIMO and distributed MIMO systems, respectively. By analyzing the disadvantages of the existing research, we present some open issues worth of studying and some promising research interest.
The rest of the paper is organized as follows. In the next section, we describe the system model of frequency synchronization in MIMO systems. Section III and Section IV address the principles and state of the art of CFO estimation and CFO compensation, respectively. And Section V concludes the paper.

SYSTEM MODEL
Consider a MIMO system with T M transmit antennas and M R receive antennas. Letting the CFO between the ith transmit antenna and the jth receive antenna is , the CFOs between the transmitter and the receiver can be expressed in
matrix form by
which is the generalized system model of frequency synchronization in MIMO systems. In many practical cases, however, some rational assumptions can be made according to specific conditions to simplify the generalized system model presented above. Therefore, the actually applied system model is usually some special case of the generalized system model.
In conventional colocated MIMO systems, since all the transmit antennas are colocated at one geographic location and all the receive antennas are colocated at another geographic location (as shown in Fig. 1), the CFOs between different transmit antennas and receive antennas can be regarded to be identical. Under this assumption, the CFO matrix reduces to

., there exists only one CFO in the system, which is similar to the case of singleantenna systems.
In distributed MIMO systems, since the transmit antennas or the receive antennas are distributed at separated geographic locations, either different oscillators are possibly used for each antenna [7], or the multipath components from each
transmit antenna may impinge on each receive antenna from different angles of arrival [8], which leads to different CFOs between each transmit antenna and each receive antenna. Therefore, there will exist multiple CFOs in the system.
In practical systems, it is usually the truth that the antennas at the base station are distributed while those at the mobile station are still colocated due to its constrained size (as shown in Fig. 2). As a result, in the downlink, the CFOs between a certain transmit antenna and all the receive antennas are identical, while the CFOs between a certain receive antenna and each transmit antenna may be different.
Under this assumption, the CFO matrix reduces to
which we will adopt in the following discussion on frequency synchronization in distributed MIMO systems.


CFO ESTIMATION IN MIMO
According to the system model given above, there is substantial difference between the CFO estimation in collocated MIMO and that in distributed MIMO. Therefore, in the following we will discuss them respectively.

CFO estimation in colocated MIMO
In colocated MIMO systems, since there exists only one CFO, which is the same as the case in singleantenna systems, the CFO estimation in colocated MIMO is a problem of scalar parameter estimation. It can be regarded as an extension of the CFO estimation in singleantenna systems and many existing CFO estimaton methods for single antenna systems can be extended to colocated MIMO systems.
The CFO estimation in colocated MIMO systems has been attracting much research interest and many CFO estimation methods can be found in the literature. [9] extended the CFO estimation method in [10] to systems with receive spatial diversity and studied the impact of spatial diversity on the performance of CFO estimation, while [11] and [12] aimed at systems with transmit spatial diversity. [13] and [14] focused on the dataaided and codeaided CFO estimation in flatfading MIMO channels. [15] applied the
CFO estimation method for flatfading MIMO channels in selective fading MIMO channels and OFDM systems. [16]~[20] addressed the CFO estimation in MIMOOFDM systems. [21] studied the CFO estimation in MIMOCDMA (code division multiple access) systems. [22]~[26] discussed the performance analysis and the design of training sequences for the CFO estimation in MIMO systems.
From the above analysis we can see that there has been adequate intensive study on the CFO estimation for co located MIMO systems and it covers a variety of aspects for CFO estimation. For example, it ranges from receive diversity to transmit diversity, from flatfading channels to selective fading channels, from CDMA systems to OFDM systems, from performance analysis to design of training sequences, and so on.

CFO estimation in distributed MIMO
As mentioned above, since there may exist multiple CFOs in distributed MIMO systems, the CFO estimation in distributed MIMO is a problem of vector parameter estimation. In addition, since the distances between each transmit antenna and each receive antenna may also be different, the propagation delays are possibly different as well, i.e., the signals transmitted from each transmit antenna to each receive antenna are not synchronous. These facts make the CFO estimation in distributed MIMO systems become much more complicated than that in colocated MIMO systems.
Considering the general case where the CFOs are possibly different for each transmit antenna, [8] addressed the joint maximum likelihood (ML) estimation of CFOs and channel gains for flatfading MIMO channels, and suggested two computationally simpler methods base on a suitable choice of the training sequence. [7] proposed two computationally efficient iterative methods based on ECM (Expectation Conditional Maximization) and SAGE (SpaceAlternating Generalized ExpectationMaximization) algorithms. [27] proposed a correlation based secondorder estimator by utilizing the orthogonality of Walsh codes, which can eliminate the multiple antenna interference. [28] extended the work in [8] to selective fading MIMO channels, considering a more general case where the CFOs are possibly different for each path, and proposed an approximate ML estimator by exploiting the crosscorrelation characteristic of the training sequences.
From the above analysis we can see that the CFO estimation for distributed MIMO systems appear incomplete. Furthermore, all the existing work only takes multiple CFOs into consideration, while still based on the assumption that the signals transmitted from each transmit antenna to each receive antenna are synchronous, which is unrealistic in practical distributed MIMO systems. Therefore, further intensive research is needed considering the generalized system model where the propagation delays between each transmit antenna and each receive antenna are possibly different.


CFO COMPENSATION IN MIMO
In colocated MIMO systems, since there exists only one CFO, we can still adopt the simple CFO correction method used in conventional singleantenna systems to perform CFO compensation prior to MIMO detection. In order to correct the CFO, the received signal at each receive antenna is multiplied by a complex exponential signal (as shown in Fig. 3), where is the CFO correction value and set equal to the estimated CFO value obtained in CFO estimation.
In distributed MIMO systems, however, since there exist multiple CFOs, the simple CFO compensation method mentioned above becomes invalid. Moreover, there is substantial difference between the CFO compensation in distributed singlecarrier MIMO systems and that in distributed MIMOOFDM systems and we will discuss them respectively.

CFO compensation in distributed singlecarrier MIMO
In distributed singlecarrier MIMO systems, in spite of the presence of multiple CFOs, the spatially multiplexed substreams will be separated with their respective CFO after MIMO detection. This gives us the opportunity to use the conventional simple CFO correction method to compensate for the CFO for each substream respectively after MIMO
detection. This operation is illustrated in Fig. 4, where , i=1,2,…..,M T is the estimated CFO value between the ith transmit antenna and the receiver.

CFO compensation in distributed MIMOOFDM
Like the singleantenna OFDM system, one of the major disadvantages of the distributed MIMOOFDM system is its sensitivity to CFOs [29], [16]. CFOs will cause a loss in the orthogonality of the subcarriers, which results in intercarrier interference (ICI) and hence significant performance degradation [16].
In distributed MIMOOFDM systems, the fast Fourier transform (FFT) is performed first to demodulate the received signal at each receive antenna prior to MIMO detection (as shown in Fig. 5). Thus, with the presence of CFOs, the ICI will occur after FFT processing and before MIMO detection. If conventional MIMO detection is performed after FFT, the ICI will occur among the separated substreams, i.e., a certain substream is affected by not only the ICI caused by itself but also the ICI introduced by other substreams. Therefore, the CFO compensation method proposed above for distributed singlecarrier MIMO systems, where the CFO for each substream is corrected respectively after MIMO detection, is not applicable to distributed MIMOOFDM systems.
In [30] and [31], some linear frequency domain equalization methods are proposed to compensate for multiple CFOs in distributed MIMOOFDM systems. Unfortunately, they involve inversion of a large matrix that is computationally complex. Although some simpler equalization methods are
also proposed, they assume small CFOs and still cannot avoid matrix inversions.
From the above analysis we can see that the problem of CFO compensation in distributed MIMOOFDM systems has been rarely addressed in the literature. The existing methods appear hard to implement, and practical lowcomplexity CFO compensation methods without need for matrix inversion still need further research.


CONCLUSION
Aiming at colocated MIMO, distributed MIMO and MIMO OFDM systems, we have addressed the principles, system model and state of the art of the estimation and compensation of CFOs, respectively. By analyzing the disadvantages of the existing research, we present some open issues worth of studying and some promising research interest. The results show that there has been adequate intensive study on the CFO estimation for colocated MIMO systems while the CFO estimation for distributed MIMO systems and the CFO compensation for MIMOOFDM systems appear incomplete and still need further research.
REFERENCES

J. G. Proakis, Digital Communications, 4th ed. New York: McGrawHill 2001.

G. J. Foschini, Layered spacetime architecture for Wireless communication in a fading environment when Using multielement antennas, Bell Labs. Tech. J., vol. 1, no. 2, pp. 4159, 1996.

J. A. C. Bingham, Multicarrier modulation for data transmission: an idea whose time has come, IEEE Cammun. Mag., vol. 28, no. 5, pp. 5 14, 1990.

G. L. StÃ¼ber, J. R. Barry, S. W. McLaughlin, Y. Li, M. A. ingram, and T G. Pratt, Broadband MIMOOFDM wireless communications, Proc IEEE, vol. 92, no. 2.

S. Zhou, M. Zhao, X. Xu, J. Wang,and Y. Yao, Distributed wireless communication systems: a new architecture for future public wireless access, IEEE Commun. Mag., vol. 41, no. 3, pp. 108113, 2003.

Z. Liu, G. Giannakis, and B. Hughes, Double differential spacetime block coding for timeselective fading channels, IEEE Trans. Commun., vol. 49, no. 9, pp. 15291539, 2001.

T.H. Pham, A. Nallanathan, and Y.C. Liang, Joint channel and frequency offset estimation in distributed MIMO flatfading channels, IEEE Trans. Wireless Commun., vol. 7, no. 2, pp. 648656, 2008.

O. Besson and P. Stoica, On parameter estimation of MIMO flatfading channels with frequency offsets, IEEE Trans. Signal Processing, vol. 51, no. 3, pp. 602 613, 2003.

M. G. Hebley and D. P. Taylor, The effect of diversity on a burstmode carrierfrequency estimator in the frequencyselective multipath channel, IEEE Trans. Commun., vol. 46, no. 4, pp. 553560, 1998.

M. Luise and R. Reggiannini, Carrier frequency recovery in alldigital modems for burstmode transmissions, IEEE Trans. Commun., vol. 43, no. 2/3/4, pp. 11691178, 1995.

E. R. Jeong, S. K. Jo, and Y. H. Lee, Least squares frequency estimation in frequencyselective channels and its application to transmissions with antenna diversity, IEEE J. Select. Areas Commun., vol. 19, no. 12, pp. 23692380, 2001.

Y.D. Kim, J. K. Lim, C.H. Suh, E.R. Jeong, and Y.
H. Lee, Carrier frequency estimation for transmissions with antenna diversity, in Proc. IEEE
VTCSpring, Birmingham, AL, 2002, pp. 15691573.

F. Sirnoens and M. Moeneclaey, A reduced complexity frequency offset estimation technique for flat fading MIMO channels, in Proc. IEEE ISCAS,
Vancouver, Canada, 2004, pp. 705708

F. Sirnoens and M. Moeneclaey, Reduced complexity dataaided and codeaided frequency offset estimation for flatfading MIMO channels, IEEE Trans. Wireless Commun., vol. 5, no. 6, pp. 15581567, 2006.

K. Lee and J. Chun, Frequencyoffset estimation for MIMO and OFDM systems using orthogonal training sequences, IEEE Trans. Veh. Technol., vol. 56, no. 1, pp. 146156, 2007.

A. N. Mody and G. L. Stuber, Synchronization for MIMO OFDM systems, in Proc. IEEE GLOBECOM, San Antonio, TX, 2001, pp. 25 29.

Y. Jiang, X. Gao, X. You, and W. Heng, Training sequence assisted frequency offset estimation for MIMO OFDM, in Proc. IEEE ICC, Istanbul, Turkey, 2006, pp. 53715376.

Y. Ogawa, K. Nishio, T. Nishimura, and T. Ohgane, Channel and frequency offset estimation for a MIMOOFDM system, in Proc. IEEE VTCFall, Los Angeles, CA, 2004, pp. 15231527 .

T. C. W. Schenk and A. van Zelst, Frequency synchronization for MIMO OFDM wireless LAN systems, in Proc. IEEE VTCFall, Orlando, FL, 2003, pp. 781785.

C. Yan, S. Li, Y. Tang, and X. Luo, Frequency synchronization in MIMO OFDM system, in Proc. IEEE VTCFall, Los Angeles, CA, 2004, pp. 1732 1734.

D.K. Hong, Y.J. Lee, and C.E. Kang, Robust frequency offset estimation for pilot symbol assisted packet CDMA with MIMO antennasystems, IEEE Commun. Lett., vol. 6, no. 6, pp. 262264, 2002.

M. Ghogho and A. Swami, Training design for multipath channel and frequencyoffset estimation in MIMO systems, IEEE Trans. Signal Processing, vol. 54, no. 10, pp. 39573965, 2006.

X. Ma, M.K. Oh, G. B. Giannakis, and D.J. Park, Hopping pilots for estimation of frequencyoffset and multiantenna channels in MIMOOFDM, IEEE Trans. Commun., vol. 53, no. 1, pp. 162172, 2005.

F. Simoens and M. Moeneclaey, Computationally efficient frequency offset estimation for flatfading MIMO channels: performance analysis and training sequence design, in Proc. IEEE GLOBECOM, Dallas, 2004, pp. 24602464.

T. Cui and C. Tellambura, Joint channel and frequency offset estimation and training sequence design for MIMO systems over frequency selective channels, in Proc. IEEE GLOBECOM, Dallas, TX, 2004, pp. 23442348.

U. Tureli, R. Ambati, D. Kivanc, and H. Liu, Performance analysis of OFDM carrier synchronization with diversity, in Proc. IEEE VTC Fall, Vancouver, Canada, 2002, pp. 3337.

Y. Yao and T.S. Ng, Correlationbased frequency offset estimation in MIMO system, in Proc. IEEE VTCFall, Orlando, FL, 2003, pp. 438 442.

S. Ahmed, S. Lambotharan, A. Jakobsson, and J. A. Chambers, MIMO frequencyselective channels with multiplefrequency offsets: estimation and detection techniques, IEE Proc.Commun., vol. 152, no. 4, pp. 489494, 2005.

T. Pollet, M. V. Bladel, and M. Moeneclaey, BER sensitivity of OFDM systems to carrier frequency offset and Weiner phase noise, IEEE Trans. Commun., vol. 43, no. 2/3/4, pp. 191193, 1995.

Z. Li, D. Qu, and G. Zhu, An equalization technique for distributed STBCOFDM system with multiple carrier frequency offsets, in Proc. IEEE WCNC, Las Vegas, NV, 2006, pp. 839843.

W. Zhang, D. Qu, and G. Zhu, Performance investigation of distributed STBCOFDM system with multiple carrier frequency offsets, in Proc. IEEE PIMRC, Helsinki, Finland, 2006, pp. 15.