Joint estimation and compensation method of RF imperfections in LTE uplink system

• US 9,537,701 B2
• Filed: 12/17/2015
• Issued: 01/03/2017
• Est. Priority Date: 03/27/2015
• Status: Expired due to Fees

• Current Assignee: Yuan Ze University
• Original Assignee: Yuan Ze University
• Inventors: Hsieh, Hung-Yang, Lin, Yi-Hsin, Huang, Sheng-Yang, Lin, Kuang-Min, Deng, Juinn-Horng, Hwang, Jeng-Kuang
• Primary Examiner(s): Washburn, Daniel
• Assistant Examiner(s): Tadese, Berhanu
• Application Number:
  US14/972,886
• Publication Number:
  US 20160285667A1
• Time in Patent Office: 383 Days
• Field of Search: 375/260
• US Class Current: 1/1
• CPC Class Codes: H04L 27/2607 Cyclic extensions
• CPC Class Codes: H04L 27/2657 Carrier synchronisation
• CPC Class Codes: H04L 27/2666 Acquisition of further OFDM parameters, e.g. bandwidth, subcarrier spacing, or guard interval length
• CPC Class Codes: H04L 27/3863 Compensation for quadrature error in the received signal

Abstract

A joint estimation and compensation method of RF imperfections in a LIE (Long Term Evolution) uplink system comprises steps: establishing a joint signal model with RF imperfections; according to the joint signal model, undertaking estimation and compensation of CFO, DC offset, multipath channel, IQ imbalance and shaping filter imbalance of a received signal; and using a frequency equalizer to equalize said received signal and determine modulation data. For reducing computational complexity, the present invention further converts the received signal from a time domain to a frequency domain to undertake frequency domain compensation. The present invention can indeed solve the problems of IQ imbalance, filter imbalance, DC offset, multipath channel and CFO and effectively estimate and compensate RF imperfections in the LTE uplink system.

Claims

1. A joint estimation and compensation method of RF imperfections in a LTE uplink system, comprising steps:

establishing a joint signal model with RF imperfections;

according to said joint signal model, undertaking an initial CFO (Carrier Frequency Offset) estimation of a received signal in a time domain so as to estimate CFO parameters and compensate said received signal;

undertaking a joint estimation of direct-c_urrent (DC) offset, multipath channel and imbalance signals of said joint signal model of said received signal, compensating said received signal in said time domain, and acquiring imbalance parameters of said imbalance signals;

determining modulation data to acquire an original transmitted signal;

determining whether error vector magnitude converges;

if yes, estimating gain mismatch parameters and phase mismatch parameters of said joint signal model according to said imbalance parameters; and

if no, using a decision-directed symbol to undertake iterative CFO estimation and compensation so as to undertake further estimation and compensation of said Imperfections until said error vector magnitude converges; and

estimating gain mismatch parameters and phase mismatch parameters of said joint signal model according to said imbalance parameters.

2. The joint estimation and compensation method according to claim 1, wherein said initial CFO estimation includes steps:

using a training sequence to estimate an initial fractional CFO so as to compensate said received signal;

estimating an initial integral CFO so as to acquire a value of said initial integral CFO; and

using a frequency estimation method to remove said CFO of said received signal.

3. The joint estimation and compensation method according to claim 2, wherein said joint signal of said received signal is expressed by

4. The joint estimation and compensation method according to claim 2, wherein said frequency estimation method converts said received signal into a post-FFT (Fast Fourier Transform) received signal and then removes said CFO from said post-FFT received signal.

5. The joint estimation and compensation method according to claim 2, wherein in said step of estimating said initial integral CFO, said integral CFO is determined via finding a maximum of

6. The joint estimation and compensation method according to claim 2, wherein said integral CFO is estimated via nulling a DC component and searching a projection of residual power of said DC offset.

7. The joint estimation and compensation method according to claim 1, wherein partial decision-directed symbols are used to undertake said iterative CFO estimation.

8. The joint estimation and compensation method according to claim 7, wherein said iterative CFO estimation is undertaken according to following equations:

9. The joint estimation and compensation method according to claim 8, wherein said decision-directed symbol ŝ

is expressed by ŝ

=ĥ

₁({tilde over (û

)}+{circumflex over (d)}1)+ĥ

₂({tilde over (û

)}*+{circumflex over (d)}*1), wherein h₁ and h₂ are imbalance parameters, u is an estimated original signal, and d is an estimated DC offset.

10. The joint estimation and compensation method according to claim 1, wherein said joint signal model of said received signal y with CFO compensation is expressed by y=h₁(ũ

+d1)+h₂(ũ

*+d*1), wherein h₁ and h₂ are imbalance coefficients involving IQ imbalance and shaping filter imbalance, and wherein d denotes DC offset, and Vector I is all N×

1 vectors, and wherein h₁ and h₂ involve a real part and an imaginary part a filter imbalance parameter h_C,I and h_C,Q, and wherein h_C,I and h_C,Q are convoluted with multipath channel h_ch and expressed by

11. The joint estimation and compensation method according to claim 10, wherein said DC offset and said multipath channel are estimated with a demodulation reference signal (DMRS), and wherein said received signal is expressed by a convolution matrix:

12. The joint estimation and compensation method according to claim 11, wherein said step of using said training sequence to estimate said DC offset and said multipath channel of said joint signal model of said received signal further includes steps:

converting said received signal into a post-FFT received signal in a frequency-domain;

multiplying said post-FFT received signal with a zero matrix to eliminate said DC offset and acquire said received signal in said frequency-domain;

using a least square and a feedback cancellation technique to estimate a reconstructed DC offset signal; and

subtracting said reconstructed DC offset signal from said received signal to obtain a residual expressed by y=h₁ũ

+h₂ũ

*.

13. The joint estimation and compensation method according to claim 1, wherein said imbalance signals include an IQ imbalance signal and a shaping filter imbalance signal.

14. The joint estimation and compensation method according to claim 13, wherein said received signal with said CFO, said DC offset, said multipath channel having been estimated and compensated is said joint signal model with imbalance signals, which is expressed by y=h₁ũ

+h₂ũ

*, and wherein h₁ and h₂ are said imbalance parameters with IQ imbalance and shaping filter imbalance, and wherein ũ

is an original signal with Δ

f=½

.

15. The joint estimation and compensation method according to claim 14, wherein a demodulation reference signal is used as said training sequence and expressed by

16. The joint estimation and compensation method according to claim 15, wherein a pseudo inverse matrix is used to estimate said imbalance parameters of said imbalance signals.

17. The joint estimation and compensation method according to claim 14, wherein said gain mismatch parameter g_T and said phase mismatch parameter φ

_T are estimated via using imbalance parameters of IQ imbalance and shaping filter imbalance of h₁ and h₂ according to an equation;

18. The joint estimation and compensation method according to claim 17, wherein a frequency-domain equalizer is used to equalize said received signal and reduce computational complexity.

19. The joint estimation and compensation method according to claim 18, wherein equalizing said received signal includes steps:

using a known matrix to compensate Δ

f=½

offset of aid received signal;

converting said received signal in a time domain into a frequency-domain received signal;

using said imbalance parameters to obtain a Δ

f=½

offset involving matrix of joint estimation of said IQ imbalance parameter and said shaping filter imbalance parameter and a joint RF effect parameter; and

multiplying said frequency-domain received signal with an inverse matrix coefficient to remove said joint RF effect parameter so as to complete an equalization process and obtain an original frequency-domain transmitted signal.

20. The joint estimation and compensation method according to claim 19, wherein after said original frequency-domain transmitted signal is obtained, an IFFT operator is used to convert said original frequency-domain transmitted signal into an original time-domain transmitted signal.