Skip to main content

Advertisement

Springer Nature Link
Log in

On the performance of a maximum likelihood method with delayed correlation for the coarse carrier frequency offset estimation of OFDM signals with multiple preamble symbols

  • Published:
Wireless Networks Aims and scope Submit manuscript

Abstract

Estimation of carrier frequency offset (CFO) is an important issue in the design of a wireless receiver that employs orthogonal frequency division multiplexing (OFDM) techniques. In this paper, using the ten short training symbols specified in the signal format of the IEEE 802.11a WLAN, we investigate the performance of a coarse CFO estimation scheme for OFDM signals with multiple preamble symbols. This scheme, which we call DC-ML, employs the maximum likelihood (ML) method with delayed correlation (DC). For AWGN channels under moderate signal to noise ratio (SNR) conditions, we develop an analysis to evaluate the variance of estimation error (VEER). The analysis is corroborated in light of simulations, and compared with the formulated Cramer-Rao lower bound (CRLB). VEER of the DC-ML in a multipath environment is studied via simulations. Numerical results show that a certain parameter combination can result in minimum VEER. Simulation results justify that the probability of estimation error (PEER) approximates Gaussian distribution in both AWGN and multipath scenarios. We also present a Two-Branch DC-ML (TBDC-ML) scheme, which comprises two correlation branches of DC-ML, and an associative ambiguity resolution algorithm. Numerical examples reveal that TBDC-ML outperforms DC-ML in both VEER and PEER. Assuming that the estimation error resulting from the two branches is jointly Gaussian, we derive the joint probability density function (pdf) and validate it via simulations.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13
Fig. 14

Similar content being viewed by others

References

  1. Van Nee, R., & Prasad, R. (2000). OFDM wireless multimedia communications. Boston: Artech House.

    Google Scholar 

  2. Bingham, J. A. C. (1990). Multicarrier modulation for data transmission: An ideal whose time has come. IEEE Communications Magazine, 28(5), 5–14. doi:10.1109/35.54342.

    Article  MathSciNet  Google Scholar 

  3. Chang, R. W. (1996). Synthesis of band-limited orthogonal signals for multi-channel data transmission. The Bell System Technical Journal, 45(10), 1775–1796.

    Google Scholar 

  4. Weinstein, S. B., & Ebert, P. M. (1971). Data transmission by frequency-division multiplexing using the discrete Fourier transform. IEEE Transactions on Communications, 19(5), 628–634. doi:10.1109/TCOM.1971.1090705.

    Article  Google Scholar 

  5. Cimini, L. J., Jr. (1985). Analysis and simulation of a digital mobile channel using orthogonal frequency division multiplexing. IEEE Transactions on Communications, 33(7), 665–675. doi:10.1109/TCOM.1985.1096357.

    Article  Google Scholar 

  6. IEEE Std 802.11a. (1999). Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications: High-speed physical layer in the 5 GHZ band.

  7. ETSI Std ETS/BRAN 30003-1. (1999). Broadband radio access networks (BRAN): HIPERLAN type 2 technical specification part I—physical layer.

  8. IEEE Std 802.16. (2004). Local and metropolitan area networks part 16: Air interface for fixed broadband wireless access systems.

  9. ETSI Std ETS 300 744. (1997). Digital video broadcasting: Framing structure, channel coding, and modulation for digital terrestrial television.

  10. Speth, M., Fechtel, S., Fock, G., & Meyr, H. (1999). Optimum receiver design for wireless broadband system using OFDM: Part I. Transactions on Communications, 47(11), 1668–1677. doi:10.1109/26.803501.

    Article  Google Scholar 

  11. Pollet, T., Van Bladel, M., & Moeneclaey, M. (1995). BER sensitivity of OFDM systems to carrier frequency offset and Wiener phase noise. IEEE Transactions on Communications, 43(2/3/4), 191–193.

    Article  Google Scholar 

  12. Steendam, H., & Moeneclaey, M. (2000). Sensitivity of orthogonal frequency-division multiplexed systems to carrier and clock synchronization errors. Signal Processing, 80(7), 1217–1229. doi:10.1016/S0165-1684(00)00031-1.

    Article  MATH  Google Scholar 

  13. Ai, B., Yang, Z. X., Pan, C. Y., Ge, J. H., & Lu, Z. (2006). On the synchronization techniques for wireless OFDM systems. IEEE Transactions on Broadcasting, 52(2), 236–244. doi:10.1109/TBC.2006.872990.

    Article  Google Scholar 

  14. Cheng, M. H., & Chou, C. C. (2006). Maximum-likelihood estimation of frequency and time offsets in OFDM systems with multiple sets of identical data. IEEE Transactions on Signal Processing, 54(7), 2848–2852. doi:10.1109/TSP.2006.874837.

    Article  Google Scholar 

  15. Van de Beck, J. J., Sandell, M., & Borjesson, P. O. (1997). ML estimation of time and frequency offset in OFDM systems. IEEE Transactions on Signal Processing, 45(7), 1800–1805. doi:10.1109/78.599949.

    Article  Google Scholar 

  16. Zheng, Z. W., Yang, Z. X., Pan, C. Y., & Zhu, Y. S. (2004). Robust synchronization and channel estimation for the OFDM-based WLAN systems. IEEE Transactions on Consumer Electronics, 50(3), 807–812. doi:10.1109/TCE.2004.1341683.

    Article  Google Scholar 

  17. Shi, K., & Serpedin, E. (2004). Coarse frame and carrier synchronization of OFDM systems: A new metric and comparison. IEEE Transactions on Wireless Communications, 3(4), 1271–1284. doi:10.1109/TWC.2004.828282.

    Article  Google Scholar 

  18. Schmidl, T. M., & Cox, D. C. (1997). Robust frequency and timing synchronization for OFDM. IEEE Transactions on Communications, 45(12), 1613–1621. doi:10.1109/26.650240.

    Article  Google Scholar 

  19. Lv, T., Li, H., & Chen, J. (2005). Joint estimation of symbol timing and carrier frequency offset of OFDM signals over fast time-varying multipath channels. IEEE Transactions on Signal Processing, 53(12), 4526–4535. doi:10.1109/TSP.2005.859233.

    Article  MathSciNet  Google Scholar 

  20. Morelli, M., & Mengali, U. (1999). An improved frequency offset estimator for OFDM applications. IEEE Communications Letters, 3(3), 75–77. doi:10.1109/4234.752907.

    Article  Google Scholar 

  21. Chang, S., & Powers, E. J. (2003). Efficient frequency-offset estimation in OFDM-based WLAN systems. IEE Electronics Letters, 39(21), 1554–1555. doi:10.1049/el:20030920.

    Article  Google Scholar 

  22. Moose, P. H. (1994). A technique for orthogonal frequency division multiplexing frequency offset correction. IEEE Transactions on Communications, 42(10), 2908–2914. doi:10.1109/26.328961.

    Article  Google Scholar 

  23. Chen, B. (2002). Maximum likelihood estimation of OFDM carrier frequency offset. IEEE Signal Processing Letters, 9(4), 123–126. doi:10.1109/97.1001648.

    Article  Google Scholar 

  24. Li, J., Liu, G., & Giannakis, G. B. (2001). Carrier frequency offset estimation for OFDM-based WLANs. IEEE Signal Processing Letters, 8(3), 80–82. doi:10.1109/97.905946.

    Article  Google Scholar 

  25. Zhang, Z., Zhao, M., Zhou, H., Liu, Y., & Gao, J. (2004). Frequency offset estimation with fast acquisition in OFDM system. IEEE Communications Letters, 8(3), 171–173. doi:10.1109/LCOMM.2004.823423.

    Article  Google Scholar 

  26. Chiavaccini, E., & Vitetta, G. M. (2004). Maximum-likelihood frequency recovery for OFDM signals transmitted over multipath fading channels. IEEE Transactions on Communications, 52(2), 244–251. doi:10.1109/TCOMM.2003.822706.

    Article  Google Scholar 

  27. Chen, H., & Pottie, G. J. (2003). A comparison of frequency offset tracking algorithms for OFDM. In Proceedings of the IEEE global communications conference, pp. 1069–1073.

  28. Classen, F., & Meyr, H. (1994). Frequency synchronization algorithms for OFDM systems suitable for communication over frequency selective fading channels. In Proceedings of the IEEE vehicular technology conference, pp. 1655–1659.

  29. Luise, M., & Reggiannini, R. (1996). Carrier frequency acquisition and tracking for OFDM systems. IEEE Transactions on Communications, 44(11), 1590–1598. doi:10.1109/26.544476.

    Article  Google Scholar 

  30. Song, H. K., You, Y. W., Paik, J. H., & Cho, Y. S. (2001). Frequency-offset synchronization and channel estimation for OFDM-based transmission. IEEE Communications Letters, 4(3), 95–97. doi:10.1109/4234.831036.

    Article  Google Scholar 

  31. Rife, D. C., & Boorstyn, A. N. (1974). Single-tone parameter estimation from discrete-time observations. IEEE Transactions on Information Theory, IT-20(5), 591–598. doi:10.1109/TIT.1974.1055282.

    Article  Google Scholar 

  32. O’Hara, B., & Patrick, A. (1999). The IEEE 802.11 handbook: A designer’s companion. New York: IEEE Press.

    Google Scholar 

  33. Leon-Garcia, A. (1994). Probability and random processes for electrical engineering. Reading: Addison-Wesley.

    Google Scholar 

Download references

Acknowledgments

This work is supported in part by the National Science Council of Taiwan under Contract No. NSC 93-2213-E-216-027, and in part by the Chung Hua University under Contract No. CHU-93-TR-003.

Author information

Authors and Affiliations

  1. Department of Electrical Engineering, Chung Hua University, No. 707, Sec. 2, WuFu Rd., Hsinchu, Taiwan, ROC

    In-Hang Chung

  2. Institute of Engineering and Science, Chung Hua University, No. 707, Sec. 2, WuFu Rd., Hsinchu, Taiwan, ROC

    Ming-Ching Yen

Authors
  1. In-Hang Chung
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Ming-Ching Yen
    View author publications

    You can also search for this author inPubMed Google Scholar

Corresponding author

Correspondence to In-Hang Chung.

Additional information

This paper was presented in part at IEEE International Conference on Communications (ICC’2005), Seoul, Korea, May 16–20, 2005.

Appendices

Appendix 1

From (5) and using the inequality, \( \left| {a + b} \right| \le \left| a \right| + \left| b \right|, \) we can obtain

$$ \begin{aligned} \frac{1}{{GN_{s} }}\left| {r\left( {G,D} \right)} \right| &\le \frac{1}{{GN_{s} }}\sum\limits_{n = 0}^{G - 1} {} \sum\limits_{m = 0}^{{N_{s} - 1}} {\left[ {\;\left| {s_{{m + nN_{s} }}^{*} \exp \left\{ { - j\left( {\frac{{2\pi \left[ {m + \left( {n + D} \right)N_{s} } \right]\Updelta k}}{N} + \theta } \right)} \right\} n_{{m + \left( {n + D} \right)N_{s} }} } \right|} \right.} \\ & + \left| {s_{{m + nN_{s} }} \exp \left\{ {j\left( {\frac{{2\pi \left[ {m + nN_{s} } \right]\Updelta k}}{N} + \theta } \right)} \right\} n_{{m + nN_{s} }}^{*} } \right|\left. { + \left| {n_{{m + nN_{s} }}^{*} n_{{m + (n + D)N_{s} }} \exp \left\{ { - j\frac{{2\pi DN_{s} \Updelta k}}{N}} \right\}} \right|\;} \right] \\ &\le \left( {R_{1} + R_{2} + R_{3} } \right)/GN_{s} , \\ \end{aligned} $$
(31)

in which \( R_{1} = \sum\limits_{n = 0}^{G - 1} {\, \sum\limits_{m = 0}^{{N_{s} - 1}} {\left| {s_{{m + nN_{s} }} } \right|\,\left| {n_{{m + (n + D)N_{s} }} } \right|} } , \) \( R_{2} = \sum\limits_{n = 0}^{G - 1} { \sum\limits_{m = 0}^{{N_{s} - 1}} {\left| {s_{{m + nN_{s} }} } \right|\,\left| {n_{{m + nN_{s} }} } \right|} } , \) and \( R_{3} = \sum\limits_{n = 0}^{G - 1} { \sum\limits_{m = 0}^{{N_{s} - 1}} {\left| {n_{{m + nN_{s} }} } \right|\,\left| {n_{{m + (n + D)N_{s} }} } \right|} } . \) Rewrite R 1 as

$$ R_{1} = \sum\limits_{n = 0}^{G - 1} {\,\left[ {\,\begin{array}{*{20}c} {\left| {s_{{nN_{s} }} } \right|} & {\left| {s_{{1 + nN_{s} }} } \right|} & \cdots & {\left| {s_{{N_{s} - 1 + nN_{s} }} } \right|} \\ \end{array} \,} \right]\;\,\left[ {\,\begin{array}{*{20}c} {\left| {n_{{(n + D)N_{s} }} } \right|} & {\left| {n_{{1 + (n + D)N_{s} }} } \right|} & \cdots & {\left| {n_{{N_{s} - 1 + (n + D)N_{s} }} } \right|} \\ \end{array} \,} \right]^{\;T} } \, = \sum\limits_{n = 0}^{G - 1} \, x_{n} \bullet y_{n} , $$
(32)

where \( x_{n} = \left[ {\,\begin{array}{*{20}c} {\left| {s_{{nN_{s} }} } \right|} & {\left| {s_{{1 + nN_{s} }} } \right|} & \cdots & {\left| {s_{{N_{s} - 1 + nN_{s} }} } \right|} \\ \end{array} \,} \right]\,, \) \( y_{n} = \left[ {\,\begin{array}{*{20}c} {\left| {n_{{(n + D)N_{s} }} } \right|} & {\left| {n_{{1 + (n + D)N_{s} }} } \right|} & \cdots & {\left| {n_{{N_{s} - 1 + (n + D)N_{s} }} } \right|} \\ \end{array} \,} \right], \) and the symbol • denotes the inner product operator. Using the Cauchy-Schwartz inequality, \( u \bullet v \le \left\| u \right\|\;\left\| v \right\|, \) in which \( \left\| {} \right\| \) represents the norm of the specified vector, we have

$$ R_{1} \le \sum\limits_{n = 0}^{G - 1} {\,\left\| {x_{n} } \right\|} \;\left\| {y_{n} } \right\| = \left[ {\,\begin{array}{*{20}c} {\left\| {x_{0} } \right\|} & {\left\| {x_{1} } \right\|} & \cdots & {\left\| {x_{G - 1} } \right\|} \\ \end{array} \,} \right]\;\left[ {\,\begin{array}{*{20}c} {\left\| {y_{0} } \right\|} & {\left\| {y_{1} } \right\|} & \cdots & {\left\| {y_{G - 1} } \right\|} \\ \end{array} \,} \right]\,^{T} . $$
(33)

Applying the Cauchy-Schwartz inequality to (33) yields

$$ R_{1} \le \left( {\sum\limits_{n = 0}^{G - 1} {\,\left\| {x_{n} } \right\|}^{2} } \right)^{1/2} \left( {\sum\limits_{n = 0}^{G - 1} {\,\left\| {y_{n} } \right\|}^{2} } \right)^{1/2} = \left( {\sum\limits_{n = 0}^{G - 1} {\,\sum\limits_{m = 0}^{{N_{s} - 1}} {\left| {s_{{m + nN_{s} }} } \right|} }^{2} } \right)^{1/2} \left( {\sum\limits_{n = 0}^{G - 1} {\,\sum\limits_{m = 0}^{{N_{s} - 1}} {\left| {n_{{m + (n + D)N_{s} }} } \right|} }^{2} } \right)^{1/2} = GN_{s} \left( {\sigma_{s}^{2} } \right)^{1/2} \left( {\sigma_{n}^{2} } \right)^{1/2} . $$
(34)

Following above procedures, we can find that \( R_{2} \le GN_{s} \left( {\sigma_{s}^{2} } \right)^{1/2} \left( {\sigma_{n}^{2} } \right)^{1/2} \) and \( R_{3} \le GN_{s} \sigma_{n}^{2} . \) Consequently,

$$ \left| {r\left( {G,D} \right)} \right|/GN_{s} \le \left( {R_{1} + R_{2} + R_{3} } \right)/GN_{s} \le 2\left( {\sigma_{s}^{2} } \right)^{1/2} \left( {\sigma_{n}^{2} } \right)^{1/2} + \sigma_{n}^{2} . $$
(35)

One may note that the value of \( \left[ {r\left( {G,D} \right)/GN_{s} } \right] + \sigma_{s}^{2} \) can be interpreted as a point lying inside a circle with radius \( 2\left( {\sigma_{s}^{2} } \right)^{1/2} \left( {\sigma_{n}^{2} } \right)^{1/2} + \sigma_{n}^{2} \) and centered at (\( \sigma_{s}^{2} , \) 0) in the complex plane, as illustrated in Fig. 15. When SNR is high, the maximum value of the real part of \( \left[ {r\left( {G,D} \right)/GN_{s} } \right] + \sigma_{s}^{2} \) will be

$$ 2\left( {\sigma_{s}^{2} } \right)^{1/2} \left( {\sigma_{n}^{2} } \right)^{1/2} + \sigma_{n}^{2} + \sigma_{s}^{2} = \sigma_{s}^{2} \left[ {2\left( {\sigma_{n}^{2} /\sigma_{s}^{2} } \right)^{1/2} + \left( {\sigma_{n}^{2} /\sigma_{s}^{2} } \right) + 1} \right] \approx \sigma_{s}^{2} . $$
(36)

Therefore,

$$ {\text{RE}}\left\{ {\left[ {r(G, D )+ \sum\limits_{n = 0}^{G - 1} {} \sum\limits_{m = 0}^{{N_{s} - 1}} {s_{{m + nN_{s} }}^{*} s_{{m + nN_{s} }} } } \right]/GN_{s} } \right\} = {\text{RE }}\left\{ { \left[ {r(G,D)/GN_{s} } \right] + \sigma_{s}^{2} } \right\} \approx \sigma_{s}^{2} . $$
(37)
Fig. 15
figure 15

Geometric interpretation of (35)

Full size image

Appendix 2

To evaluate (11), we note that \( {\text{IM }}\left[ {r\left( {G,D} \right)} \right] = \left[ {r\left( {G,D} \right) - r^{*} \left( {G,D} \right)} \right]/(2j). \) Also note that the \( n_{{m + nN_{s} }}^{*} n_{{m + (n + D)N_{s} }} \) term in (5) will not take effect when expectation is taken. Consequently,

$$ \begin{aligned} {{\text {IM}}} \left[ {r\left( {G,D} \right)} \right] = \frac{1}{2j}\sum\limits_{n = 0}^{G - 1} {\sum\limits_{m = 0}^{{N_{s} - 1}} {\left[ {s_{{m + nN_{s} }}^{*} \exp \left\{ { - j\left[ {\frac{{2\pi \left( {m + nN_{s} } \right) \Updelta k}}{N} + \theta } \right]} \right\}\left( {n_{{m + \left( {n + D} \right)N_{s} }} \exp \left\{ { - j\frac{{2\pi DN_{s} \Updelta k}}{N}} \right\} - n_{{m + nN_{s} }} } \right)} \right.} } \\ + s_{{m + nN_{s} }} \exp \left\{ {j\left[ {\frac{{2\pi \left( {m + nN_{s} } \right) \Updelta k}}{N} + \theta } \right]} \right\}\left( {n_{{m + nN_{s} }}^{*} - n_{{m + \left( {n + D} \right)N_{s} }}^{*} \exp \left\{ {j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right) \\ \left. { + \left( {n_{{m + nN_{s} }}^{*} n_{{m + (n + D)N_{s} }} \exp \left\{ { - j\frac{{2\pi DN_{s} \Updelta k}}{N}} \right\} - n_{{m + nN_{s} }} n_{{m + \left( {n + D} \right)N_{s} }}^{*} \exp \left\{ {j\frac{{2\pi DN_{s} \Updelta k}}{N}} \right\}} \right)} \right]. \\ \end{aligned} $$
(38)

Using (38), we can obtain \( E\left\{ {\left[ {{\text{ IM }}\left\{ {r (G,D)} \right\} } \right]^{2} } \right\} \) as expressed in (39).

$$ \begin{aligned} E\left\{ {\left[ {{\text{ IM }}\left\{ {r \left( {G,D} \right)} \right\} } \right]^{2} } \right\} = \frac{1}{4}\left\{ {\sum\limits_{n = 0}^{G - 1} {\sum\limits_{m = 0}^{{N_{s} - 1}} {{\kern 1pt} \sum\limits_{i = 0}^{{G - 1}} {\sum\limits_{l = 0}^{{N_{s} - 1}} { \left[ {E \left\{ {s_{{m + nN_{s} }}^{*} s_{{l + iN_{s} }} } \right\} E \left( {n_{{m + \left( {n + D} \right)N_{s} }} n_{{l + \left( {i + D} \right)N_{s} }}^{*} - n_{{m + \left( {n + D} \right)N_{s} }} n_{{l + iN_{s} }}^{*} \exp \left\{ { - j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right.} \right.} } } } } \right. \\ \left. {\left. { + n_{{m + nN_{s} }} n_{{l + iN_{s} }}^{*} - n_{{m + nN_{s} }} n_{{l + \left( {i + D} \right)N_{s} }}^{*} \exp \left\{ {j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right)} \right] \\ + \left[ {E \left\{ {s_{{m + nN_{s} }} s_{{l + iN_{s} }}^{*} } \right\} E \left( {n_{{m + nN_{s} }}^{*} n_{{l + iN_{s} }} - n_{{m + nN_{s} }}^{*} n_{{l + \left( {i + D} \right)N_{s} }} \exp \left\{ { - j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right.} \right. \\ + n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{l + \left( {i + D} \right)N_{s} }} \left. {\left. { - n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{l + iN_{s} }} \exp \left\{ {j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right) } \right] \\ + \left[ {E \left\{ {s_{{m + nN_{s} }}^{*} s_{{l + iN_{s} }}^{*} } \right\} \exp \left\{ { - j\left( {\frac{{2\pi (m + nN_{s} + l + iN_{s} )\Updelta k}}{N} + 2\theta } \right)} \right\}E \left( {n_{{m + \left( {n + D} \right)N_{s} }} n_{{l + iN_{s} }} \exp \left\{ { - j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right.} \right. \\ \left. {\left. { + n_{{m + nN_{s} }} n_{{l + \left( {i + D} \right)N_{s} }} \exp \left\{ { - j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\} - n_{{m + \left( {n + D} \right)N_{s} }} n_{{l + \left( {i + D} \right)N_{s} }} \exp \left\{ { - j\left( {\frac{{4\pi DN_{s} \Updelta k}}{N}} \right)} \right\} - n_{{m + nN_{s} }} n_{{l + iN_{s} }} } \right)} \right] \\ + \left[ {E \left\{ {s_{{m + nN_{s} }} s_{{l + iN_{s} }} } \right\} \exp \left\{ {j\left( {\frac{{2\pi (m + nN_{s} + l + iN_{s} )\Updelta k}}{N} + 2\theta } \right)} \right\}E \left( {n_{{m + nN_{s} }}^{*} n_{{l + \left( {i + D} \right)N_{s} }}^{*} \exp \left\{ {j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right.} \right. \\ \left. {\left. { + n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{l + iN_{s} }}^{*} \exp \left\{ {j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\} - n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{l + \left( {i + D} \right)N_{s} }}^{*} \exp \left\{ {j\left( {\frac{{4\pi DN_{s} \Updelta k}}{N}} \right)} \right\} - n_{{m + nN_{s} }}^{*} n_{{l + iN_{s} }}^{*} } \right)} \right] \\ + \left[ {E \left( {n_{{m + nN_{s} }}^{*} n_{{m + \left( {n + D} \right)N_{s} }} n_{{l + iN_{s} }} n_{{l + \left( {i + D} \right)N_{s} }}^{*} - n_{{m + nN_{s} }}^{*} n_{{m + \left( {n + D} \right)N_{s} }} n_{{l + iN_{s} }}^{*} n_{{l + \left( {i + D} \right)N_{s} }} \exp \left\{ { - j\frac{{4\pi DN_{s} \Updelta k}}{N}} \right\}} \right.} \right. \\ + n_{{m + nN_{s} }} n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{l + iN_{s} }}^{*} n_{{l + \left( {i + D} \right)N_{s} }} \left. {\left. { - n_{{m + nN_{s} }} n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{l + iN_{s} }} n_{{l + \left( {i + D} \right)N_{s} }}^{*} \exp \left\{ {j\frac{{4\pi DN_{s} \Updelta k}}{N}} \right\}} \right)} \right] \\ \left. { + E \left[ {\text{terms which contain triple products of noise or its complex conjugate}} \right]} \right\} \\ \end{aligned}$$
(39)

For all values of m, n, l and i in (39), \( E \left\{ {s_{{m + nN_{s} }} s_{{l + iN_{s} }} } \right\}, \) \( E \left\{ {n_{{m + nN_{s} }}^{*} n_{{m + \left( {n + D} \right)N_{s} }} n_{{l + iN_{s} }}^{*} n_{{l + \left( {i + D} \right)N_{s} }} } \right\}, \) \( E\left\{ {s_{{m + nN_{s} }}^{*} s_{{l + iN_{s} }}^{*} } \right\}, \) and \( E\,\left\{ {n_{{m + nN_{s} }} n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{l + iN_{s} }} n_{{l + \left( {i + D} \right)N_{s} }}^{*} } \right\} \)are zero. If \( m \ne l, \) \( E \left\{ {s_{{m + nN_{s} }}^{*} s_{{l + iN_{s} }} } \right\} = 0 \) and \( E \left\{ {s_{{m + nN_{s} }} s_{{l + iN_{s} }}^{*} } \right\} = 0 \) for all n and i. In addition, since an AWGN channel is considered, \( E \left\{ {n_{y}^{{}} n_{z}^{{}} } \right\} \)=\( E \left\{ {n_{y}^{ *} n_{z}^{*} } \right\} = 0 \) for \( y = m + nN_{s} \) or \( y = m + \left( {n + D} \right)N_{s} \), and \( z = l + iN_{s} \) or \( z = l + \left( {i + D} \right)N_{s} . \) Moreover, the last term in (39) is zero. Consequently, (39) can be simplified to

$$ \begin{aligned} E\left\{ {\left[ {{\text{ IM }}\left\{ {r \left( {G,D} \right)} \right\} } \right]^{2} } \right\} = \frac{1}{4}\left\{ {\sum\limits_{n = 0}^{G - 1} { \sum\limits_{{{\text{i}} = 0}}^{G - 1} {} \sum\limits_{m = 0}^{{N_{s} - 1}} { \left[ {E \left\{ {s_{{m + nN_{s} }}^{*} s_{{m + iN_{s} }} } \right\} E \left( {n_{{m + \left( {n + D} \right)N_{s} }} n_{{m + \left( {i + D} \right)N_{s} }}^{*} } \right.} \right.} } } \right. \\ \left. {\left. { - n_{{m + \left( {n + D} \right)N_{s} }} n_{{m + iN_{s} }}^{*} \exp \left\{ { - j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\} + n_{{m + nN_{s} }} n_{{m + iN_{s} }}^{*} - n_{{m + nN_{s} }} n_{{m + \left( {i + D} \right)N_{s} }}^{*} \exp \left\{ {j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right)} \right] \\ + \left[ {E \left\{ {s_{{m + nN_{s} }} s_{{m + iN_{s} }}^{*} } \right\} E \left( {n_{{m + nN_{s} }}^{*} n_{{m + iN_{s} }} - n_{{m + nN_{s} }}^{*} n_{{m + (i + D)N_{s} }} \exp \left\{ { - j(\frac{{2\pi DN_{s} \Updelta k}}{N})} \right\}} \right.} \right. \\ + n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{m + \left( {i + D} \right)N_{s} }} \left. {\left. { - n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{m + iN_{s} }} \exp \left\{ {j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right) } \right] \\ + \left[ {E \left( {n_{{m + nN_{s} }}^{*} n_{{m + \left( {n + D} \right)N_{s} }} n_{{m + iN_{s} }} n_{{m + \left( {i + D} \right)N_{s} }}^{*} } \right.} \right.\left. {\left. {\left. { + n_{{m + nN_{s} }} n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{m + iN_{s} }}^{*} n_{{m + \left( {i + D} \right)N_{s} }} } \right){\kern 1pt} } \right]{\kern 1pt} {\kern 1pt} } \right\}. \\ \end{aligned} $$
(40)

In (40), expectation for products of the noise terms, \( E \left\{ {n_{{m + nN_{s} }}^{*} n_{{m + iN_{s} }} } \right\}, \) \( E \left\{ {n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{m + \left( {i + D} \right)N_{s} }} } \right\}, \) \( E \left\{ {n_{{m + nN_{s} }}^{*} n_{{m\, + \,\left( {n + D} \right)N_{s} }} n_{{m + iN_{s} }} n_{{m + \left( {i + D} \right)N_{s} }}^{*} } \right\}, \) \( E \left\{ {n_{{m + nN_{s} }} n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{m + iN_{s} }}^{*} n_{{m + \left( {i + D} \right)N_{s} }} } \right\}, \) \( E \left\{ {n_{{m + nN_{s} }} n_{{m + iN_{s} }}^{*} } \right\}, \) and \( E \left\{ {n_{{m + \left( {n + D} \right)N_{s} }} n_{{m + \left( {i + D} \right)N_{s} }}^{*} } \right\} \) will be nonzero only when \( n = i \). But values of the other terms, \( E \left\{ {n_{{m + \left( {n + D} \right)N_{s} }} n_{{m + iN_{s} }}^{*} } \right\} , \) \( E \left\{ {n_{{m + nN_{s} }} n_{{m + \left( {i + D} \right)N_{s} }}^{*} } \right\} , \) \( E \left\{ {n_{{m + nN_{s} }}^{*} n_{{m + \left( {i + D} \right)N_{s} }} } \right\} , \) and \( E \left\{ {n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{m + iN_{s} }} } \right\} \) will depend on whether D ≥ G or D < G. In case D ≥ G, all these terms will be zero. Consequently,

$$ \begin{aligned} E\left\{ {\left[ {{\text{ IM }}\left\{ {r \left( {G,D} \right)} \right\} } \right]^{2} } \right\} &= \frac{1}{4}\left\{ {\sum\limits_{n = 0}^{G - 1} { \sum\limits_{m = 0}^{{N_{s} - 1}} { \left[ {E \left\{ {s_{{m + nN_{s} }}^{*} s_{{m + nN_{s} }} } \right\} E \left( {n_{{m + \left( {n + D} \right)N_{s} }} n_{{m + \left( {n + D} \right)N_{s} }}^{*} } \right.} \right.} } } \right.\left. {\left. { + n_{{m + nN_{s} }} n_{{m + nN_{s} }}^{*} } \right)} \right] \\ & \quad + \left[ {E \left\{ {s_{{m + nN_{s} }} s_{{m + iN_{s} }}^{*} } \right\} E \left( {n_{{m + nN_{s} }}^{*} n_{{m + nN_{s} }} } \right.} \right.\left. {\left. { + n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{m + \left( {n + D} \right)N_{s} }} } \right) } \right] \\ & \quad + \left[ {E \left( {n_{{m + nN_{s} }}^{*} n_{{m + \left( {n + D} \right)N_{s} }} n_{{m + nN_{s} }} n_{{m + \left( {n + D} \right)N_{s} }}^{*} } \right.} \right.\left. {\left. {\left. { + n_{{m + nN_{s} }} n_{{m + \left( {n + D} \right)N_{s} }}^{*} n_{{m + nN_{s} }}^{*} n_{{m + \left( {n + D} \right)N_{s} }} } \right){\kern 1pt} } \right]\,} \right\} \\ & = GN_{s} \left( {\sigma_{s}^{2} \sigma_{n}^{2} + \sigma_{n}^{4} / 2} \right) .\\ \end{aligned} $$
(41)

On the other hand, as D < G, some of the above terms will not be zero for n + D = i or i + D = n. Thus,

$$ \begin{aligned} E\left\{ {\left[ { IM \left\{ {r \left( {G,D} \right)} \right\} } \right]^{{{\kern 1pt} 2}} } \right\} & = \frac{1}{4}\left\{ {\sum\limits_{n = 0}^{G - 1} { \sum\limits_{m = 0}^{{N_{s} - 1}} { \left[ {E \left\{ {s_{{m + nN_{s} }}^{*} s_{{m + nN_{s} }} } \right\} E \left( {n_{{m + (n + D)N_{s} }} n_{{m + (n + D)N_{s} }}^{*} } \right.} \right.} } } \right.\left. {\left. { + n_{{m + nN_{s} }} n_{{m + nN_{s} }}^{*} } \right)} \right] \\ & + \left[ {E \left\{ {s_{{m + nN_{s} }} s_{{m + iN_{s} }}^{*} } \right\} E \left( {n_{{m + nN_{s} }}^{*} n_{{m + nN_{s} }} } \right.} \right.\left. {\left. { + n_{{m + (n + D)N_{s} }}^{*} n_{{m + (n + D)N_{s} }} } \right) } \right] \\ & + \left[ {E \left( {n_{{m + nN_{s} }}^{*} n_{{m + (n + D)N_{s} }} n_{{m + nN_{s} }} n_{{m + (n + D)N_{s} }}^{*} } \right.} \right.\left. {\left. {\left. { + n_{{m + nN_{s} }} n_{{m + (n + D)N_{s} }}^{*} n_{{m + nN_{s} }}^{*} n_{{m + (n + D)N_{s} }} } \right)} \right]{\kern 1pt} } \right\} \\ & - \frac{1}{4}\left\{ {\sum\limits_{n = 0}^{G - D - 1} {\sum\limits_{m = 0}^{{N_{s} - 1}} {} \left[ {E\left\{ {s_{{m + nN_{s} }}^{*} s_{{m + (n + D)N_{s} }} } \right\}E\left\{ {n_{{m + (n + D)N_{s} }} n_{{m + (n + D)N_{s} }}^{*} } \right\} \exp \left\{ { - j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right.} } \right. \\ & + \left. {E\left\{ {s_{{m + nN_{s} }} s_{{m + (n + D)N_{s} }}^{*} } \right\} E\left\{ {n_{{m + (n + D)N_{s} }}^{*} n_{{m + (n + D)N_{s} }} } \right\} \exp \left\{ {j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right] \\ & + \sum\limits_{i = 0}^{G - D - 1} {\sum\limits_{m = 0}^{{N_{s} - 1}} {} \left[ {E \left\{ {s_{{m + (i + D)N_{s} }}^{*} s_{{m + iN_{s} }} } \right\} E \left\{ {n_{{m + (i + D)N_{s} }} n_{{m + (i + D)N_{s} }}^{*} } \right\}\exp \left\{ {j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right.} \\ & \left. {\left. { + E \left\{ {s_{{m + (i + D)N_{s} }} s_{{m + iN_{s} }}^{*} } \right\}E\left\{ {n_{{m + (i + D)N_{s} }}^{*} n_{{m + (i + D)N_{s} }} } \right\} \exp \left\{ { - j\left( {\frac{{2\pi DN_{s} \Updelta k}}{N}} \right)} \right\}} \right] } \right\} \\ & = GN_{s} \sigma_{s}^{2} \sigma_{n}^{2} + GN_{s} \sigma_{n}^{4} /2 - (G - D)N_{s} \sigma_{s}^{2} \sigma_{n}^{2} \cos \left( {2\pi DN_{s} \Updelta k/N} \right) \\ & \approx DN_{s} \sigma_{s}^{2} \sigma_{n}^{2} + GN_{s} \sigma_{n}^{4} /2,\,if\, 2\pi DN_{s} \Updelta k/N \ll 1. \\ \end{aligned} $$
(42)

Appendix 3

Note that Δk 1 = Δk 2 = Δk. Using (11) and (38) with (D, G) replaced by \( \left( {D_{1} , G_{1} } \right) \) or \( (D_{2} , G_{2} ), \) we find

$$ \begin{aligned} E\left\{ {\Updelta \tilde{k}_{1} \Updelta \tilde{k}_{2} } \right\} & = \frac{{ - N^{2} }}{{16\pi^{2} D_{1} G_{1} D_{2} G_{2} N_{s}^{4} \sigma_{s}^{4} }}E\left\{ {\sum\limits_{n = 0}^{{G_{1} - 1}} {} \sum\limits_{m = 0}^{{N_{s} - 1}} {\left[ {s_{{m + nN_{s} }}^{*} \exp \left\{ { - j\left[ {\frac{{2\pi (m + nN_{s} ) \Updelta k}}{N} + \theta } \right]} \right\}^{ } } \right.} } \right. \\ & \quad \times \left( {n_{{m + (n + D_{1} )N_{s} }} \exp \left\{ { - j\frac{{2\pi D_{1} N_{s} \Updelta k}}{N}} \right\} - n_{{m + nN_{s} }} } \right) \\ \,\; & \quad + s_{{m + nN_{s} }} \exp \left\{ {j\left[ {\frac{{2\pi (m + nN_{s} ) \Updelta k}}{N} + \theta } \right]} \right\}\left( {n_{{m + nN_{s} }}^{*} - n_{{m + (n + D_{1} )N_{s} }}^{*} \exp \left\{ {j\frac{{2\pi D_{1} N_{s} \Updelta k}}{N}} \right\}} \right) \\ & & \;\;{\kern 1pt} \;\left. { + \left( {n_{{m + nN_{s} }}^{*} n_{{m + (n + D_{1} )N_{s} }} \exp \left\{ { - j\frac{{2\pi D_{1} N_{s} \Updelta k}}{N}} \right\} - n_{{m + nN_{s} }} n_{{m + (n + D_{1} )N_{s} }}^{*} \exp \left\{ {j\frac{{2\pi D_{1} N_{s} \Updelta k}}{N}} \right\}} \right)} \right] \\ & \quad \times \sum\limits_{i = 0}^{{G_{2} - 1}} {} \sum\limits_{l = 0}^{{N_{s} - 1}} {\left[ {s_{{l + iN_{s} }}^{*} \exp \left\{ { - j\left[\frac{{2\pi (l + iN_{s} ) \Updelta k}}{N} + \theta \right]} \right\}\left( {n_{{l + (i + D_{2} )N_{s} }} \exp \left\{ { - j(\frac{{2\pi D_{2} N_{s} \Updelta k}}{N})} \right\} - n_{{l + iN_{s} }} } \right)} \right.} \\ & \;\;\,{\kern 1pt} + s_{{l + iN_{s} }} \exp \left\{ {j\left[ {\frac{{2\pi (l + iN_{s} ) \Updelta k}}{N} + \theta } \right]} \right\}\left( {n_{{l + iN_{s} }}^{*} - n_{{l + (i + D_{2} )N_{s} }}^{*} \exp \left\{ {j\left( {\frac{{2\pi D_{2} N_{s} \Updelta k}}{N}} \right)} \right\}} \right) \\ & \;\;\,\left. {\left. { + \left( {n_{{l + iN_{s} }}^{*} n_{{l + (i + D_{2} )N_{s} }} \exp \left\{ { - j\frac{{2\pi D_{2} N_{s} \Updelta k}}{N}} \right\} - n_{{l + iN_{s} }} n_{{l + (i + D_{2} )N_{s} }}^{*} \exp \left\{ {j\frac{{2\pi D_{2} N_{s} \Updelta k}}{N}} \right\}} \right)} \right] } \right\}. \\ \end{aligned} $$
(43)

Let p denote the value of the expectation \( E\left\{ \cdot \right\} \) on the right hand side of (43). Note that if \( m \ne l, \) \( E \left\{ {s_{{m + nN_{s} }}^{*} s_{{l + iN_{s} }} } \right\} = 0 \) and \( E \left\{ {s_{{m + nN_{s} }} s_{{l + iN_{s} }}^{*} } \right\} = 0 \) for all n and i. Also note that \( D_{1} < D_{2} \). After simplification and some mathematical manipulations, p and \( E\left\{ {\Updelta \tilde{k}_{1} \Updelta \tilde{k}_{2} } \right\} \)can be obtained as

$$ p = N_{s} \left[ {\sum\limits_{n = 0}^{{G_{2} - 1}} { \left( { - 2\sigma_{s}^{2} \sigma_{n}^{2} } \right) + } \sum\limits_{n = 0}^{{G_{2} - D_{1} - 1}} { \left( {2\sigma_{s}^{2} \sigma_{n}^{2} } \right) + } \sum\limits_{i = 0}^{{G_{1} - D_{2} - 1}} { \left( {2\sigma_{s}^{2} \sigma_{n}^{2} } \right) + } \sum\limits_{i = 0}^{{G_{2} - 1}} { \left( { - 2\sigma_{s}^{2} \sigma_{n}^{2} } \right) } } \right] = - 4D_{1} N_{s} \sigma_{s}^{2} \sigma_{n}^{2} . $$
(44)
$$ E\left\{ {\Updelta \tilde{k}_{1} \Updelta \tilde{k}_{2} } \right\} = \frac{{N^{2} \sigma_{n}^{2} }}{{4\pi^{2} D_{2} G_{1} G_{2} N_{s}^{3} \sigma_{s}^{2} }} = \frac{{N^{2} }}{{4\pi^{2} D_{2} G_{1} G_{2} N_{s}^{3} }} \cdot \frac{1}{\text{SNR}}. $$
(45)

Rights and permissions

Reprints and permissions

About this article

Cite this article

Chung, IH., Yen, MC. On the performance of a maximum likelihood method with delayed correlation for the coarse carrier frequency offset estimation of OFDM signals with multiple preamble symbols. Wireless Netw 16, 641–657 (2010). https://doi.org/10.1007/s11276-008-0159-5

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11276-008-0159-5

Keywords