A semi-blind channel estimation method for broadband OFDM satellite systems

Abstract

This invention discloses a semi-blind channel estimation method for broadband OFDM satellite systems. It involves evaluating OFDM frequency domain data and selecting subcarriers that simultaneously meet two preset criteria to make them suitable for the MPSK frequency offset blind estimation algorithm. Frequency offset blind estimation parameters are set, and frequency offset estimates for a set of subcarriers are obtained according to the MPSK algorithm. These estimates are then analyzed to obtain a set of reference values ​​for subsequent frequency offset interpolation. Linear interpolation of these reference values ​​yields the frequency offsets for all subcarriers, which are then used for frequency offset compensation. This invention can estimate the phase rotation superimposed on different OFDM symbols on the same subcarrier due to Doppler frequency shift. Using block pilot-based channel estimation, it compensates for amplitude variations and phase shifts caused by fading on different subcarriers, achieving high-performance channel estimation results.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a semi-blind channel estimation method for broadband OFDM satellite systems. Background Technology

[0002] Orthogonal Frequency Division Multiplexing (OFDM) technology, due to its high spectral efficiency and ability to effectively combat interference between signal waveforms, has been widely applied in mobile broadband satellite communication systems as the demand for global roaming and high-speed broadband increases. However, the influence of actual satellite channels can cause severe distortion of OFDM signals at the receiver. Therefore, employing channel estimation techniques at the receiver to compensate for the effects of channel fading is crucial for OFDM systems. In existing satellite communication systems, channel estimation is typically based on pilot symbols transmitted by the transmitter, but this method requires a large number of pilot symbols, reducing spectral efficiency. To address this, a semi-blind channel estimation method has been proposed, which estimates channel coefficients using only a small number of pilot symbols and the characteristics of the received signal.

[0003] Existing semi-blind channel estimation algorithms can be broadly categorized into four types: The first type utilizes nonlinear filtering methods for time-domain semi-blind estimation. The second type employs the Expectation-Maximum (EM) algorithm based on the maximum likelihood criterion and its improved versions. This type of algorithm avoids the matrix inversion problem encountered in traditional semi-blind estimation through iterative reception, but its computational complexity is high. The third type uses a small amount of pilot information to eliminate phase uncertainty in blind estimation, but this method requires a large number of OFDM symbols for statistical averaging. The fourth type is the neural network-based semi-blind estimation method that has emerged in recent years. While this method boasts outstanding nonlinear mapping performance, it also introduces limitations in generalization ability. In summary, although there are numerous semi-blind channel estimation algorithms at present, most are developed based on blind channel estimation and suffer from a series of problems such as high computational complexity, the need for a large amount of observation data, slow convergence speed, and propagation distortion.

[0004] Furthermore, for broadband OFDM satellite communication systems, increased signal bandwidth implies increased subcarrier spacing. Assuming a constant channel coherence bandwidth, increased signal bandwidth leads to greater differences in frequency-selective fading experienced by different subcarriers. Moreover, as shown by the Doppler shift formula, different carrier frequencies result in different Doppler shifts, and the frequency offset difference increases with the carrier frequency difference, becoming even more pronounced with high-speed satellite movement. Therefore, the fading and frequency offset differences between subcarriers in broadband OFDM satellite systems are greater than in non-broadband systems. On the other hand, if the pilots in the time domain are sparsely inserted, the difficulty of channel estimation in such systems is further increased. Therefore, designing a low-complexity, semi-blind channel estimation technique applicable to broadband OFDM satellite systems has become a crucial issue. Summary of the Invention

[0005] To address the limitations and defects of existing technologies, this invention provides a semi-blind channel estimation method for broadband OFDM satellite systems, comprising:

[0006] Step S1: Perform a Fast Fourier Transform on the OFDM signal;

[0007] Step S2: Judge the OFDM frequency domain data and select subcarriers that simultaneously meet two preset criteria so that the subcarriers are suitable for the MPSK frequency offset blind estimation algorithm;

[0008] Step S3: Set the frequency offset blind estimation parameters;

[0009] Step S4: Obtain a set of frequency offset estimates {f} on subcarriers according to the MPSK frequency offset blind estimation algorithm. n};

[0010] Step S5: Estimate the frequency offset of this group of values ​​{f} n} Analysis is performed to obtain a set of reference values ​​{f} for subsequent frequency offset interpolation. n ′};

[0011] Step S6: For this set of reference values ​​{f n Linear interpolation is performed on the subcarriers to obtain the frequency offset. The frequency offset Used for frequency offset compensation.

[0012] Optional, also includes:

[0013] Estimate the amplitude variation and phase rotation caused by channel fading based on known block pilot signals;

[0014] The amplitude variation and the phase rotation are compensated.

[0015] Optionally, the subcarriers that simultaneously satisfy two preset criteria include:

[0016] The MPSK symbols on the subcarrier can be acquired at equal intervals;

[0017] The number of MPSK symbols that can be acquired at equal intervals is greater than or equal to a preset value.

[0018] Optionally, step S3 includes:

[0019] The frequency resolution of the Fast Fourier Transform is obtained by the following expression:

[0020]

[0021] Where, N FFT f is the number of points in the Fast Fourier Transform. s The sampling rate of the MPSK symbols on the subcarrier;

[0022] The frequency offset between two adjacent blind frequency offset estimation subcarriers is obtained using the following expression:

[0023]

[0024] Where v is the relative velocity between the transmitter and receiver, c is the speed of light, and f is the speed of light. c Δk is the overall carrier frequency of the OFDM signal, f0 is the interval between two adjacent blind frequency offset estimated subcarriers, α is the interval between the subcarriers, and α is the relative angle between the transmitter and the receiver.

[0025] The frequency resolution of the Fast Fourier Transform is set to be less than the frequency offset between two adjacent blind frequency offset estimation subcarriers, as expressed in the following expression:

[0026]

[0027] The expression obtained is as follows:

[0028]

[0029] Where r = f0 / f s , which represents the ratio of the subcarrier spacing to the sampling rate of the MPSK symbols on the subcarrier.

[0030] Optionally, step S5 includes:

[0031] Among a group of subcarriers with the same frequency offset estimate, the frequency offset estimate with the middle subcarrier number is selected as the reference for subsequent frequency offset interpolation.

[0032] Obtain a set of reference values ​​{f} for frequency offset interpolation n ′} and its corresponding subcarrier number.

[0033] The present invention has the following beneficial effects:

[0034] This invention provides a semi-blind channel estimation method for broadband OFDM satellite systems. It involves evaluating OFDM frequency domain data and selecting subcarriers that simultaneously meet two preset criteria, making these subcarriers suitable for the MPSK frequency offset blind estimation algorithm. Frequency offset blind estimation parameters are set, and frequency offset estimates for a set of subcarriers are obtained according to the MPSK algorithm. These estimates are then analyzed to obtain a set of reference values ​​for subsequent frequency offset interpolation. Linear interpolation of these reference values ​​yields the frequency offsets for all subcarriers, which are then used for frequency offset compensation. This invention can estimate the phase rotation superimposed by Doppler frequency shift for different OFDM symbols on the same subcarrier. Using block pilot-based channel estimation, it compensates for amplitude variations and phase shifts caused by fading on different subcarriers. Therefore, the semi-blind channel estimation method for broadband OFDM satellite systems provided by this invention can achieve high-performance channel estimation results. Attached Figure Description

[0035] Figure 1 The above is a time-frequency diagram of a broadband OFDM satellite system with time-domain sparse pilots provided in Embodiment 1 of the present invention.

[0036] Figure 2 The flowchart illustrates the semi-blind channel estimation method for a broadband OFDM satellite system provided in Embodiment 1 of the present invention.

[0037] Figure 3 This is a schematic diagram of the picket fence effect in FFT frequency offset estimation provided in Embodiment 1 of the present invention.

[0038] Figure 4 This is a schematic diagram comparing the bit error rates of different channel estimation schemes provided in Embodiment 1 of the present invention.

[0039] Figure 5a This is the first constellation diagram of the 510th subcarrier when EbN0 = 25dB, provided in Embodiment 1 of the present invention.

[0040] Figure 5b This is a second constellation diagram of the 510th subcarrier when EbN0 = 25dB, provided in Embodiment 1 of the present invention.

[0041] Figure 5c This is the third constellation diagram of the 510th subcarrier when EbN0 = 25dB, provided in Embodiment 1 of the present invention.

[0042] Figure 5d This is the fourth constellation diagram of the 510th subcarrier when EbN0 = 25dB, provided in Embodiment 1 of the present invention.

[0043] Figure 5eThis is the fifth constellation diagram of the 510th subcarrier when EbN0 = 25dB, provided in Embodiment 1 of the present invention.

[0044] Figure 6a This is the first constellation diagram of the 100th subcarrier when EbN0 = 25dB, provided in Embodiment 1 of the present invention.

[0045] Figure 6b This is a second constellation diagram of the 100th subcarrier when EbN0 = 25dB, provided in Embodiment 1 of the present invention.

[0046] Figure 6c This is the third constellation diagram of the 100th subcarrier when EbN0 = 25dB, provided in Embodiment 1 of the present invention.

[0047] Figure 6d This is the fourth constellation diagram of the 100th subcarrier when EbN0 = 25dB, provided in Embodiment 1 of the present invention.

[0048] Figure 6e This is the fifth constellation diagram of the 100th subcarrier when EbN0 = 25dB, provided in Embodiment 1 of the present invention. Detailed Implementation

[0049] To enable those skilled in the art to better understand the technical solution of the present invention, the semi-blind channel estimation method for broadband OFDM satellite systems provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0050] Example 1

[0051] 1. Overview

[0052] To address the aforementioned issues with broadband orthogonal frequency division multiplexing (OFDM) satellite systems featuring time-domain sparse pilots...

[0053] To address the channel estimation problem, this embodiment proposes a semi-blind estimation method that combines conventional channel estimation based on pilots with blind frequency offset estimation to obtain high-performance channel estimation results.

[0054] 2. Broadband OFDM Semi-Blind Channel Estimation Technology Scheme

[0055] Small-scale fading in satellite channels is generally classified into two types: frequency-selective fading caused by multipath effects and time-selective fading caused by Doppler effects. In pilot-based channel estimation techniques, comb pilots are typically used for estimating time-selective fading channels, while block pilots are used for estimating frequency-selective fading channels. For broadband OFDM satellite systems with time-domain sparse pilots, such as... Figure 1For example, if the pilot signals are densely distributed in the frequency domain at a certain moment, they can effectively compensate for the frequency-selective fading on each subcarrier at that moment. However, compensation for time-selective fading cannot be achieved using sparse pilot signals in the time domain.

[0056] To address the aforementioned problems, this embodiment proposes a semi-blind channel estimation method for broadband OFDM satellite systems with time-domain sparse pilots. The basic flow of the semi-blind channel estimation method for broadband OFDM satellite systems provided in this embodiment is as follows: Figure 2 As shown, the specific steps are as follows:

[0057] 1) Judge the frequency domain OFDM data received after Fast Fourier Transform (FFT) and select subcarriers that can be used for the frequency offset blind estimation algorithm of Multiple Phase Shift Keying (MPSK).

[0058] 2) Set the frequency offset blind estimation parameters according to the actual system scenario;

[0059] 3) Using the MPSK frequency offset blind estimation algorithm, a set of frequency offset values ​​{f} on the subcarriers are obtained. n};

[0060] 4) The obtained frequency offset estimates {f n} Analyze and determine a set of reference values ​​{f} for subsequent frequency offset interpolation. n ′};

[0061] 5) Set the reference {f n The frequency offset on all subcarriers is obtained through linear interpolation. Therefore, compensation will be provided;

[0062] 6) Using known block pilots, estimate the amplitude changes and phase rotations caused by channel fading, and perform compensation processing.

[0063] The following section will introduce the innovative aspects of the semi-blind channel estimation method for broadband OFDM satellite systems provided in this embodiment, namely... Figure 2 The main components include: frequency offset blind estimation subcarrier selection, frequency offset blind estimation parameter setting, analysis and determination of frequency offset interpolation benchmark and frequency offset interpolation, etc.

[0064] 2.1 Frequency Offset Blind Estimation Subcarrier Selection

[0065] In OFDM satellite systems, a subcarrier may have different modulation schemes in different symbols, so subcarriers need to be screened before frequency offset blind estimation.

[0066] In OFDM satellite systems, each subcarrier can be considered a single-carrier signal. If the MPSK symbols on a subcarrier cannot be acquired at equal intervals, it means that the MPSK symbol sampling rate of that single-carrier signal is not fixed. Therefore, it is impossible to use FFT operations to obtain accurate spectral information during frequency offset blind estimation. Secondly, there should be a sufficient number of MPSK symbols that can be acquired at equal intervals. If there are insufficient MPSK symbols, the number of signal points for FFT operations in frequency offset blind estimation will be insufficient, leading to spectral distortion and making it impossible to accurately estimate the frequency offset value of the subcarrier. This embodiment extracts subcarriers that simultaneously meet the above two criteria so that a relatively accurate frequency offset information can be extracted using an FFT-based MPSK frequency offset blind estimation algorithm.

[0067] Although the MPSK frequency offset blind estimation algorithm based on FFT can obtain relatively good frequency offset estimates, the frequency offset estimated by these subcarriers still has a certain error. This error originates from the picket fence effect inherent in FFT itself. The error caused by the picket fence effect is mainly divided into two types: First, the frequency resolution of FFT is insufficient, causing the frequency offset estimation results of multiple neighboring subcarriers to fall into the same picket fence, that is, the frequency offset estimation results of multiple neighboring subcarriers are the same; second, when the Doppler frequency offset of the subcarrier is not equal to an integer multiple of the FFT spectral spacing, the estimation error caused by simply using the spectral line with the largest amplitude as the frequency offset estimate.

[0068] like Figure 3 As shown in the figure, f r For the frequency resolution of the FFT, Δf d The frequency offset between two adjacent subcarriers is given by f. r The frequency points spaced at intervals are the fence of the FFT, with f as the interval. d The frequency points indicated by the dashed lines represent the frequency offset values ​​on each subcarrier. Figure 3 As can be seen, the frequency point indicated by the dashed line is the midpoint between two adjacent left and right fences, and also the dividing point of the FFT estimation results. When the frequency offset is (k-1)Δf d 、kΔf d (k+1)Δf d All three subcarriers fall within the frequency point Kf r When the fence is within the range, the two errors mentioned above occur, meaning that the frequency offset value estimated after FFT is Kf. r Furthermore, the further the actual frequency deviates from the target fence frequency, the greater the estimation error.

[0069] To address the first type of error, the next section will provide a reasonable setting for the parameters of the frequency offset blind estimation. The second type of error will be addressed in Section 2.3.

[0070] 2.2 Frequency-biased blind estimation parameter settings

[0071] Let the number of points in the FFT be N. FFTThe sampling rate of the MPSK symbol on the subcarrier is f s Then the frequency resolution of the FFT is:

[0072]

[0073] Because the frequency offsets differ across subcarriers in an OFDM satellite system, estimating the frequency offsets of different subcarriers separately requires that the frequency resolution of the FFT be smaller than the frequency offset between two adjacent blindly estimated subcarriers. In an OFDM satellite system, this frequency offset is:

[0074]

[0075] In the formula, f0 is the subcarrier spacing, and Δk is the estimated subcarrier spacing between two adjacent blind frequency offsets. Therefore:

[0076]

[0077] Based on expression (2-3), we obtain the following expression:

[0078]

[0079] Where r = f0 / f s , representing the ratio of subcarrier spacing to MPSK symbol sampling rate. Therefore, the receiver needs to select an appropriate number of FFT points N based on the relative motion velocities v and α between the transmitter and receiver. FFT The parameters are: the number of subcarrier intervals Δk for blind frequency offset estimation, and the ratio r of the subcarrier interval to the MPSK symbol sampling rate. However, in practice, due to limitations such as hardware equipment and signal length, the parameter settings cannot guarantee that the frequency resolution of all FFTs is less than the frequency offset between two adjacent blind frequency offset estimation subcarriers. Therefore, it is still necessary to use the characteristics of the second type of error for further correction.

[0080] 2.3 Analysis and Determination of Frequency Offset Interpolation Reference

[0081] As analyzed in Section 2.1 regarding the picket fence effect, for subcarriers whose frequency offsets fall within the same picket fence, the subcarrier with the more central its number, the closer its true frequency offset is to the target picket fence's frequency. For example... Figure 3 As shown, the frequency offset is (k-1)Δf d 、kΔf d (k+1)Δf d The FFT frequency offset estimates for all three subcarriers are Kf. r Clearly, among these three subcarriers, the actual frequency offset on the second subcarrier is closest to the estimated value. Therefore, to address the second problem mentioned in Section 2.1, among a group of subcarriers with the same estimated frequency offset, the frequency offset value in the middle of the subcarrier numbers should be taken as the reference for subsequent frequency offset interpolation.

[0082] 2.4 Frequency Offset Interpolation

[0083] In OFDM satellite systems, although different carrier frequencies on different subcarriers lead to different frequency offsets, the frequency offsets on adjacent subcarriers are equal because the subcarrier spacing is constant. Inspired by this, the semi-blind channel estimation method for broadband OFDM satellite systems provided in this embodiment utilizes this property to interpolate and estimate the frequency offsets on different subcarriers. A set of reference values ​​{f} for frequency offset interpolation is obtained. n After determining the frequency offset of ′} and its corresponding subcarrier number, linear interpolation is performed to obtain the frequency offset estimates for all subcarriers. Subsequently, frequency offset compensation is performed using this method to complete the OFDM subcarrier frequency offset blind estimation module in the broadband OFDM satellite system semi-blind channel estimation method provided in this embodiment.

[0084] 3. Simulation Results

[0085] This section presents simulation results to verify the effectiveness of the proposed broadband OFDM semi-blind channel estimation technique, comparing its performance with several traditional methods. The comparison schemes include: 1) Non-blind channel estimation: channel estimation for all OFDM symbols based solely on block pilots; 2) Fully blind channel estimation: full-blind equalization processing for each data subcarrier of the OFDM signal; 3) Ideal channel estimation: assuming the receiver knows all channel coefficients. This simulation sets the OFDM signal bandwidth to 200MHz, the number of subcarriers to 1024, and the modulation scheme to QPSK. Regarding Doppler frequency offset, it is assumed that the center frequency of the receiver baseband signal is 0MHz (i.e., the overall carrier frequency offset has been compensated during the frequency offset synchronization stage), and the difference in Doppler frequency offset between two adjacent subcarriers is 4Hz. The block pilot is inserted only at the first OFDM symbol.

[0086] Figure 4 A comparison of the system bit error rate using different channel estimation schemes is presented. Simulation results show that traditional non-blind and fully blind channel estimation methods cannot effectively compensate for channel errors under the simulation conditions. However, the semi-blind channel estimation scheme proposed in this embodiment successfully completes the channel estimation task.

[0087] When EbN0 = 25dB Figures 5a-5e The constellation diagrams of subcarrier No. 510 after compensation under different channel estimation schemes were compared. Because this subcarrier is very close to the center frequency, the overall frequency offset correction effect is better during the frequency offset synchronization phase. Figures 5a-5e As can be seen, the channel estimation results of different channel estimation schemes on the intermediate subcarriers all accurately restored the phase information of the signal; except for the fully blind estimation, they all accurately restored the amplitude information of the signal.

[0088] When EbN0 = 25dB Figures 6a-6e The constellation diagrams of subcarrier number 100 after compensation under different channel estimation schemes were compared. Because this subcarrier is far from the center frequency, the initial overall frequency offset correction failed to adequately correct the frequency offset on this subcarrier. From... Figures 6a-6e As can be seen from the simulation, under these conditions, neither non-blind nor fully blind estimation can estimate the phase of subcarriers with residual frequency offsets and compensate for them; the final constellation diagram remains a loop. However, the semi-blind channel estimation scheme proposed in this embodiment can estimate the frequency offsets on different subcarriers, accurately compensate for both the amplitude and phase of the received signal, and successfully completes the channel estimation task.

[0089] In summary, traditional non-blind channel estimation, like fully blind channel estimation, cannot address the problem of large frequency offset differences among different subcarriers in broadband OFDM signals when pilot signals are scarce. The semi-blind channel estimation scheme proposed in this embodiment, however, effectively solves this problem.

[0090] 4. Conclusion

[0091] This embodiment provides a semi-blind channel estimation method for a broadband OFDM satellite system. The method includes modules for single-carrier MPSK frequency offset blind estimation, OFDM subcarrier frequency offset blind estimation, and OFDM subcarrier frequency offset interpolation. It can estimate the phase rotation superimposed on different OFDM symbols on the same subcarrier due to Doppler frequency shift. Subsequently, channel estimation based on block pilots is used to compensate for amplitude variations and phase shifts caused by fading on different subcarriers. Finally, this embodiment designs a set of comparative experiments to verify the effectiveness of the proposed semi-blind channel estimation scheme.

[0092] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of the present invention, and the present invention is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A semi-blind channel estimation method for a broadband OFDM satellite system, characterized in that, include: Step S1: Perform a Fast Fourier Transform on the OFDM signal; Step S2: Judge the OFDM frequency domain data and select subcarriers that simultaneously meet two preset criteria so that the subcarriers are suitable for the MPSK frequency offset blind estimation algorithm; Step S3: Set the frequency offset blind estimation parameters; Step S4: Obtain a set of frequency offset estimates on subcarriers using the MPSK frequency offset blind estimation algorithm. ; Step S5: Estimate the frequency offset for this group. The analysis was performed to obtain a set of reference values ​​for subsequent frequency offset interpolation. ; Step S6: For this set of reference values Perform linear interpolation to obtain the frequency offset on all subcarriers. The frequency offset Used for frequency offset compensation; The subcarriers that simultaneously satisfy two preset criteria include: The MPSK symbols on the subcarrier can be acquired at equal intervals; The number of MPSK symbols that can be acquired at equal intervals is greater than or equal to a preset value; Step S3 includes: The frequency resolution of the Fast Fourier Transform is obtained by the following expression: Equation (2-1) in, N FFT The number of points in the Fast Fourier Transform. The sampling rate of the MPSK symbols on the subcarrier; The frequency offset between two adjacent blind frequency offset estimation subcarriers is obtained using the following expression: Equation (2-2) in, v The relative velocity between the transmitter and receiver. At the speed of light, The overall carrier frequency of the OFDM signal. Estimate the number of subcarrier intervals for two adjacent blind frequency offsets. The interval of the subcarriers, The relative angle between the transmitter and receiver; The frequency resolution of the Fast Fourier Transform is set to be less than the frequency offset between two adjacent blind frequency offset estimation subcarriers, as expressed in the following expression: Equation (2-3) The expression obtained is as follows: Equation (2-4) in, , which represents the ratio of the subcarrier spacing to the sampling rate of the MPSK symbols on the subcarrier.

2. The semi-blind channel estimation method for broadband OFDM satellite systems according to claim 1, characterized in that, Also includes: Estimate the amplitude variation and phase rotation caused by channel fading based on known block pilot signals; The amplitude variation and the phase rotation are compensated.

3. The semi-blind channel estimation method for broadband OFDM satellite systems according to claim 1, characterized in that, Step S5 includes: Among a group of subcarriers with the same frequency offset estimate, the frequency offset estimate with the middle subcarrier number is selected as the reference for subsequent frequency offset interpolation. Obtain a set of reference values ​​for frequency offset interpolation and its corresponding subcarrier number.

Images