A frequency fine estimation method of high dynamic narrowband satellite signal

Abstract

The application discloses a method for frequency fine estimation of high-dynamic narrow-band satellite signals, which comprises the following steps: performing frequency estimation on the front and rear synchronization codes to obtain a basic frequency deviation; performing multi-path frequency deviation estimation within a preset frequency deviation range; and finally selecting a frequency estimation value from the multi-path frequency deviation estimation values. The application can simply and accurately estimate the frequency deviation of the high-dynamic narrow-band satellite signals by using the existing front and rear synchronization codes in the burst, significantly improves the frequency estimation precision, and is suitable for the application of the narrow-band signals in the high-dynamic environment with a large Doppler frequency deviation and a certain Doppler frequency variation rate.

Description

Technical Field

[0001] This invention relates to the field of satellite communication technology, and in particular to a method for precise frequency estimation of high dynamic narrowband satellite signals. Background Technology

[0002] In MF-TDMA satellite communication systems, transmitted information is processed according to the TDMA mechanism, with bursts as the smallest unit. Frequency deviations and Doppler shifts between transmitting and receiving equipment result in a certain frequency offset in the received baseband signal, severely impacting signal reception performance. Therefore, the ability to quickly and accurately estimate and calibrate this frequency offset is a prerequisite for correct data reception and an essential step in satellite communication systems.

[0003] In an MF-TDMA system, after the system completes acquisition and synchronization, it corrects the timing and frequency of receiving other bursts based on the frequency offset and time information of the acquired synchronization burst. This allows the system to receive other bursts with small frequency deviations at a predetermined time after exiting the acquisition mode. Typically, the receiver uses a phase-locked loop to adjust the frequency of the local carrier to achieve carrier synchronization between transmitted and received data. However, the short burst communication mechanism of the MF-TDMA system is not suitable for using a tracking loop. TDMA systems often perform fine frequency estimation after coarse frequency synchronization to further remove small frequency offsets in the received burst signals, so that the signal phase deviation is within the acceptable range of subsequent modules.

[0004] Furthermore, in satellite communication systems, if the user terminal (such as various types of aircraft) and the central station have a high relative velocity and relative acceleration, it will generate a large Doppler frequency offset and Doppler rate of change, resulting in a large frequency offset in the received burst signals, which seriously affects the signal reception performance. This situation is particularly serious with narrowband signals. Therefore, how to quickly and accurately estimate and calibrate the frequency offset of narrowband signals in a high-dynamic environment is a prerequisite for correct data reception and also a relatively difficult aspect of narrowband satellite communication systems.

[0005] Commonly used algorithms for frequency estimation include M&M, Kay, Fitz, and L&R. All four algorithms can reach the Cramer-Rao bound at high signal-to-noise ratios (SNRs), but they have the following drawbacks for fine frequency estimation: The M&M algorithm has a large correction range and a low threshold (approximately 0 dB), making its implementation relatively complex and suitable for coarse frequency offset estimation; the Kay algorithm has a large correction range and is relatively simple to implement, but its threshold is very high (approximately 9 dB), making it unsuitable for low to medium SNRs. To work at these SNRs, the signal must be accumulated before applying the Kay algorithm, requiring a large number of known symbols and demanding high precision in burst structure calculations; the Fitz and L&R algorithms have relatively small correction ranges, similar performance, and no clear performance boundary, making them suitable only for fine frequency estimation. However, as the SNR decreases, the root mean square error (RMSE) of these algorithms deviates further from the Cramer-Rao bound, and correlation calculations are required, making implementation more complex.

[0006] Patent 1 (Joint Frequency Estimation Method of Synchronization Codes and Pilots, ZL 2020 1 0893666.X) and Patent 2 (Frequency Estimation Method of Pilots and Synchronization Codes, ZL 2020 1 0893682.9) are frequency estimation methods that use synchronization codes and pilots together. These methods have certain requirements for signal frequency deviation, requiring the signal frequency deviation to be less than a certain range. However, in high dynamic range narrowband signal systems, acceleration is more sensitive to frequency deviation, and the frequency deviation is often large, which limits the application of this algorithm.

[0007] In summary, existing algorithms require a large number of consecutive known symbols for precise frequency estimation. The more known symbols there are, the better the frequency estimation performance, but the more complex the calculation process becomes. For bursts with a small number of consecutive known symbols, the existing algorithms have poor performance in precise frequency estimation and are more computationally complex. For high-dynamic narrowband signals, the impact of frequency deviation is greater, and algorithms with simpler calculations cannot meet the requirements. Summary of the Invention

[0008] The purpose of this invention is to provide a method for precise frequency estimation of high dynamic narrowband satellite signals in order to solve the above-mentioned problems, comprising the following steps:

[0009] Step 1: Use the pre- and post-synchronization codes to estimate the fundamental frequency deviation Δf;

[0010] Step 2: Perform multi-channel frequency deviation estimation within the preset range of frequency deviation to obtain the possible values ​​of multi-channel frequency deviation estimation Δfd;

[0011] Step 3: Select the estimated frequency deviation df from the possible values ​​of the multi-channel frequency deviation estimate;

[0012] in,

[0013] Step 2, obtaining the possible values ​​of the multi-channel frequency offset estimate, includes the following steps:

[0014] Step 2.1: Determine the preset range of frequency deviation [-f0, f0], where f0 is the preset range threshold value of frequency deviation, which is selected based on the expected residual frequency deviation of the system;

[0015] Step 2.2: Calculate the number of multiple frequencies N within the preset frequency deviation range, N = floor(f0×Ts×Ls), where Ts is the symbol period of the burst, Ls is the number of interval symbols between the center symbols of the preamble and the center symbols of the postamble selected for the burst, and floor() represents rounding down;

[0016] Step 2.3: Calculate the possible value Δfd of the multi-channel frequency deviation estimate, Δfd(k) = Δf + k / (Ts×Ls), where k ∈{-N,-N+1,…,N-1,N};

[0017] Step 3, selecting the estimated frequency deviation from the possible values ​​of the multi-channel frequency deviation estimate, includes the following steps:

[0018] Step 3.1: Perform frequency offset removal operation on signal X according to the possible values ​​Δfd estimated by the multi-channel frequency offset to obtain the frequency offset-removed signal X. ’ X ’ (k) = X×e(-j×2Π×Δfd(k)×t), where k ∈{-N,-N+1,…,N-1,N},t=0:Ts:(L_b-1)×Ts, and L_b is the burst length of the signal;

[0019] Step 3.2: Calculate the power value P of the "known signal" in the multi-channel de-estimation signal, P(k) = X ’ _k(k)×conj(X_k), where k ∈{-N,-N+1,…,N-1,N}, X_k represents a known symbol in signal X, X ’ _k represents the signal X after frequency offset estimation. ’ In the known symbols, conj() represents conjugate computation;

[0020] Step 3.3: Select the frequency deviation of the path with the largest power value as the estimated frequency deviation df, df = Δfd(k0), where P(k0) = max{P(k)}, k ∈{-N,-N+1,…,N-1,N}.

[0021] The present invention achieves the above objectives through the following technical solution: For burst signals with pre- and post-synchronization codes, the present invention uses the frequency estimation of the pre- and post-synchronization codes as the basic frequency deviation, performs multi-channel frequency deviation estimation within a preset range of frequency deviation, and selects the channel with the largest signal power after removing the frequency deviation as the actual estimated frequency deviation value.

[0022] The beneficial effects of this invention are as follows: In the case of high dynamic narrowband satellite signals, this invention can easily and accurately estimate the frequency deviation by using the existing pre- and post-synchronization codes in the burst, which significantly improves the accuracy of frequency estimation and is suitable for high dynamic environments with large Doppler frequency deviation and a certain Doppler change rate. Attached Figure Description

[0023] Figure 1 This is a typical schematic diagram of a sudden structure.

[0024] Figure 2 This is a typical schematic diagram of a burst structure containing a pilot block.

[0025] Figure 3 This is a flowchart illustrating the working principle of the present invention.

[0026] Figure 4 This is a flowchart illustrating the working principle of the pre- and post-synchronization code basic frequency offset estimation method.

[0027] Figure 5 This is a flowchart illustrating the working principle of multi-channel frequency offset estimation within a preset frequency offset range.

[0028] Figure 6 This is a flowchart illustrating the working principle of selecting the estimated frequency deviation from multiple paths.

[0029] Figure 7 The waveform is reference waveform 1 in Table A-1 of Appendix A of the DVB-RCS2 standard low-layer protocol (ETSI EN 301 545-2). The frequency estimation performance of the present invention and the previous version is compared in the same frequency offset range.

[0030] Figure 8 The frequency estimation performance of the present invention and the previous one is compared under the same frequency offset range and Doppler change rate, based on reference waveform No. 1 in Table A-1 of Appendix A of the DVB-RCS2 standard low-layer protocol (ETSI EN 301 545-2).

[0031] Figure 9 The waveform is reference waveform 1 in Table A-1 of Appendix A of the DVB-RCS2 standard low-layer protocol (ETSI EN 301 545-2). The frequency estimation performance of this invention is compared and simulated in different frequency offset ranges.

[0032] Figure 10The waveform is reference waveform 1 in Table A-1 of Appendix A of the DVB-RCS2 standard low-layer protocol (ETSI EN 301 545-2). The frequency estimation performance of this invention is compared and simulated at different Doppler change rates.

[0033] In the figures: A1 is the performance simulation curve of the reference waveform in the frequency offset range of [-15, +15] Hz using the basic frequency offset estimation; A2 is the performance simulation curve of the reference waveform in the frequency offset range of [-15, +15] Hz using the frequency offset estimation of this invention; B1 is the performance simulation curve of the reference waveform in the frequency offset range of [-15, +15] Hz under the condition of 100 Hz / s Doppler rate of change using the basic frequency offset estimation; B2 is the performance simulation curve of the reference waveform in the frequency offset range of [-15, +15] Hz under the condition of 100 Hz / s Doppler rate of change using the frequency offset estimation of this invention. Simulation curves; C1, C2, C3, and C4 are the performance simulation curves of the reference waveform using the frequency offset estimation of this invention in the frequency offset range of [-700, +700] Hz, [-500, +500] Hz, [-300, +300] Hz, and [-100, +100] Hz, respectively; D1, D2, D3, and D4 are the performance simulation curves of the reference waveform using the frequency offset estimation of this invention in the frequency offset range of [-100, +100] Hz under the Doppler rate of change conditions of 10 Hz / s, 20 Hz / s, 50 Hz / s, and 100 Hz / s, respectively. Detailed Implementation

[0034] The invention will be further described below with reference to the accompanying drawings: as shown in the drawings Figure 1 The diagram shows a typical burst structure. In an MF-TDMA satellite communication system, the forward and reverse link bursts employ a general burst structure with linear modulation. Each burst contains one or more user payload segments, while other segments are known symbols, such as preamble and postamble codes. Preamble and postamble codes are typically present in every burst, and pilot blocks may also be present. (See attached diagram.) Figure 2 The diagram shows a typical burst structure with pilot blocks, in which the pilot blocks are evenly distributed among the payload segments of uniform size.

[0035] As attached Figure 3 As shown, the present invention relates to a method for precise frequency estimation of high dynamic narrowband satellite signals, comprising the following steps:

[0036] Step 1: Use the pre- and post-synchronization codes to estimate the fundamental frequency deviation Δf. The specific process is shown in the attached figure. Figure 4 As shown:

[0037] Step 1.1: Calculate the center symbol angle of the preamble: The known symbol of the selected preamble is q, and the receiver receives the selected preamble symbol as q. ’The selected preamble symbol number is Lq, and the selected preamble center symbol angle is... , where arg{} is the angle calculation function;

[0038] Step 1.2: Calculate the center symbol angle of the postsynchronization code: The known symbol of the postsynchronization code is h, and the postsynchronization code symbol received by the receiver is h. ’ The number of symbols in the postsynchronization code is Lh, and the center symbol angle of the postsynchronization code is... ;

[0039] Step 1.3: Calculate the frequency deviation: Ln is the number of symbols between the first symbol of the selected preamble and the last symbol of the postamble, Ls is the number of symbols between the center symbol of the selected preamble and the center symbol of the postamble, Ls = Ln - Lq / 2 - Lh / 2, Ts is the symbol period, and rem{} is the remainder function. Then the fundamental frequency deviation is calculated as follows: ;

[0040] Step 2: Within the preset range of frequency deviation, perform multi-channel frequency deviation estimation to obtain the possible values ​​Δfd for multi-channel frequency deviation estimation. The specific process is shown in the attached figure. Figure 5 As shown:

[0041] Step 2.1: Determine the preset range of frequency deviation [-f0, f0], where f0 is the preset range threshold value of frequency deviation, which is selected based on the expected residual frequency deviation of the system;

[0042] Step 2.2: Calculate the number of multiple frequencies N within the preset frequency deviation range, N = floor(f0×Ts×Ls), where Ts is the symbol period of the burst, Ls is the number of interval symbols between the center symbols of the preamble and the center symbols of the postamble selected for the burst, and floor() represents rounding down;

[0043] Step 2.3: Calculate the possible value Δfd of the multi-channel frequency deviation estimate, Δfd(k) = Δf + k / (Ts×Ls), where k ∈{-N,-N+1,…,N-1,N};

[0044] Step 3: Select the estimated frequency deviation df from the possible values ​​of the multi-channel frequency deviation estimate. The specific process is shown in the appendix. Figure 6 As shown:

[0045] Step 3.1: Perform frequency offset removal operation on signal X according to the possible values ​​Δfd estimated by the multi-channel frequency offset, to obtain the frequency offset-removed signal X. ’ X ’(k) = X×e(-j×2Π×Δfd(k)×t), where k ∈{-N,-N+1,…,N-1,N},t=0:Ts:(L_b-1)×Ts, and L_b is the burst length of the signal;

[0046] Step 3.2: Calculate the power value P of the "known signal" in the multiplexed signal after frequency offset estimation, specifically, as shown in the appendix. Figure 2 As shown, the power value of "preamble + postamble + pilot signal" is calculated here as P(k) = P_q(k) + P_h(k) + P_d(k), where P_q represents the power value of the "preamble", P_h represents the power value of the "postamble", P_d represents the power value of the "pilot signal", and P_q(k) = q ’ (k) ×conj(q(k)), P_h(k)= h ’ (k) ×conj(h(k)), P_d(k) = d ’ (k)×conj(d(k)); q(k), h(k), and d(k) represent the preamble, postamble, and pilot symbols in the signal, respectively. ’ (k), h ’ (k), d ’ (k) represent the preamble, postamble, and pilot symbols in the signal after frequency offset estimation, respectively, k ∈{-N,-N+1,…,N-1,N}, and conj() represents conjugate calculation;

[0047] Step 3.3: Select the frequency deviation of the path with the largest power value as the estimated frequency deviation df, df = Δfd(k0), where P(k0) = max{P(k)}, k ∈{-N,-N+1,…,N-1,N}.

[0048] The simulation was performed using reference wave 1 with a total burst length of 664 symbols from Table A-1 in Appendix A of the DVB-RCS2 standard low-layer protocol (ETSI EN 301 545-2). The number of pre- and post-synchronization symbols was 27, and the number of pilot symbols was 26. It was assumed that the symbol rate R of the narrowband signal was 20Ksps, the phase deviation was random within the range of [-1 / 2π, +1 / 2π], and the number of Monte Carlo simulations was 2000. The horizontal axis represents the signal-to-noise ratio in dB, and the vertical axis represents the normalized standard deviation of the frequency offset.

[0049] Figure 7This is a performance simulation comparison chart of the present invention and the basic frequency offset estimation (Patent 1: Joint Frequency Estimation Method for Pre- and Post-Synchronization Codes, ZL 2020 1 0893666.X) when the reference waveform is in the frequency offset range of [-0.75‰, +0.75‰]×R ([-15, +15]Hz). Curve A1 is the performance simulation curve using the basic frequency offset estimation, and curve A2 is the performance simulation curve using the frequency offset estimation of the present invention. Figure 7 It can be seen that under the frequency offset condition of [-0.75‰, +0.75‰]×R, when the signal-to-noise ratio is less than 8dB, the performance index of the normalized frequency offset standard deviation is significantly improved compared with the basic frequency offset estimation algorithm when the signal-to-noise ratio is less than 8dB. When the signal-to-noise ratio is 0dB, the normalized frequency offset standard deviation is reduced from the original 1.5×10 -4 Upgraded to 4×10 -5 Additionally, when the normalized frequency deviation standard deviation is less than 8 × 10⁻⁶ -4 The received signal can be recovered well afterward. Therefore, the basic frequency offset estimation algorithm needs a signal-to-noise ratio greater than 4dB to work properly, while this invention can work properly at a signal-to-noise ratio as low as -5dB.

[0050] Figure 8 This is a performance simulation comparison chart of the present invention and the basic frequency offset estimation (Patent 1: Joint Frequency Estimation Method for Synchronization Codes before and after Synchronization Codes, ZL 2020 1 0893666.X) with the reference waveform in the frequency offset range of [-0.75‰, +0.75‰]×R ([-15, +15]Hz), plus a Doppler rate of change of 100Hz / s. Curve B1 is the performance simulation curve using the basic frequency offset estimation; B2 is the performance simulation curve using the frequency offset estimation of the present invention. Figure 8 It can be seen that, under the conditions of [-0.75‰, +0.75‰]×R frequency offset and 100Hz / s Doppler rate of change, the normalized frequency offset standard deviation of the basic frequency offset estimation algorithm is much greater than 8×10. -4 The basic frequency offset estimation algorithm cannot function properly under these conditions; however, when the signal-to-noise ratio is greater than 0 dB, the normalized frequency offset standard deviation performance index of this invention can reach 8 × 10⁻⁶. -5 It works normally.

[0051] Figure 9 This is a simulation comparison of the frequency estimation performance of the present invention using the reference waveform at different frequency offset ranges. Curves C1, C2, C3, and C4 are the normalized frequency offset standard deviation curves at different signal-to-noise ratios when the frequency offset is [-3.5%, +3.5%]×R ([-700, +700]Hz), [-2.5%, +2.5%]×R ([-500, +500]Hz), [-1.5%, +1.5%]×R ([-300, +300]Hz), and [-0.5%, +0.5%]×R ([-100, +100]Hz). Figure 9 It can be seen that when the frequency offset range is [-0.5%, +0.5%]×R, the normalized frequency offset standard deviation can reach 4×10⁻⁶ when the signal-to-noise ratio is 0dB. -5 As the preset frequency offset range increases, the normalized frequency offset standard deviation increases overall, and the frequency offset estimation performance deteriorates.

[0052] Figure 10 This is a simulation comparison of the frequency estimation performance of this invention using the reference waveform within the frequency offset range of [-0.5%, +0.5%]×R and at different Doppler rates of change. Curves D1, D2, D3, and D4 are the normalized frequency offset standard deviation curves for Doppler rates of change of 10Hz / s, 20Hz / s, 50Hz / s, and 100Hz / s, respectively, at different signal-to-noise ratios. Figure 10 It can be seen that when the signal-to-noise ratio is greater than 0 dB, the normalized frequency offset standard deviation is less than 8 × 10⁻⁶. -5 As the preset Doppler rate of change increases, the normalized frequency offset standard deviation increases overall, and the frequency offset estimation performance deteriorates.

[0053] Therefore, under high dynamic conditions, the present invention can greatly improve the frequency estimation range and accuracy of narrowband signals, thereby improving the quality of the received signal.

[0054] This invention targets burst signals with pre- and post-synchronization codes. It uses the frequency estimation of the pre- and post-synchronization codes as the basic frequency deviation, performs multi-channel frequency deviation estimation within a preset range of frequency deviation, and selects the channel with the largest signal power after removing the frequency deviation as the actual estimated frequency deviation value.

[0055] This invention enables the accurate estimation of frequency offset using existing pre- and post-synchronization codes in high-dynamic narrowband signal conditions, requiring only a small number of known symbols. It significantly improves the accuracy of frequency estimation and is suitable for improving narrowband signal reception quality in high-dynamic environments with large frequency offsets or a certain Doppler variation rate.

[0056] The technical solutions of the present invention are not limited to the specific embodiments described above. Any technical modifications made in accordance with the technical solutions of the present invention fall within the protection scope of the present invention.

Claims

1. A method for precise frequency estimation of high dynamic narrowband satellite signals, comprising the following steps: Step 1: Use the pre- and post-synchronization codes to estimate the fundamental frequency deviation Δf; Step 2: Perform multi-channel frequency deviation estimation within the preset range of frequency deviation to obtain the possible values ​​of multi-channel frequency deviation estimation Δfd; Step 3: Select the estimated frequency deviation df from the possible values ​​of the multi-channel frequency deviation estimate; in, Step 2, obtaining the possible values ​​of the multi-channel frequency offset estimate, includes the following steps: Step 2.1: Determine the preset range of frequency deviation [-f0, f0], where f0 is the preset range threshold value of frequency deviation, which is selected based on the expected residual frequency deviation of the system; Step 2.2: Calculate the number of multiple frequencies N within the preset range of frequency deviation, N = floor(f0×Ts×Ls), where Ts is the symbol period of the burst, Ls is the number of interval symbols between the center symbol of the preamble and the center symbol of the postamble selected in the burst, and floor() represents rounding down; Step 2.3: Calculate the possible value Δfd of the multi-channel frequency deviation estimate, Δfd(k) = Δf + k / (Ts×Ls), where k∈{-N,-N+1,…,N-1,N}; Step 3, selecting the estimated frequency deviation from the possible values ​​of the multi-channel frequency deviation estimate, includes the following steps: Step 3.1: Perform frequency offset removal operation on signal X according to the possible values ​​Δfd estimated by the multi-channel frequency offset, to obtain the frequency offset-removed signal X. ’ X ’ (k) = X×e(-j×2Π×Δfd(k)×t), where k ∈{-N,-N+1,…,N-1,N},t=0:Ts:(L_b-1)×Ts, and L_b is the burst length of the signal; Step 3.2: Calculate the power value P of the "known signal" in the multi-channel de-estimation signal, P(k) = X ’ _k(k)×conj(X_k(k)), where k ∈{-N,-N+1,…,N-1,N}, X_k represents a known symbol in signal X, X ’ _k represents the signal X after frequency offset estimation. ’ In the known symbols, conj() represents conjugate computation; Step 3.3: Select the frequency deviation of the path with the largest power value as the estimated frequency deviation df, df = Δfd(k0), where P(k0) = max{P(k)}, k ∈{-N,-N+1,…,N-1,N}.

2. The method according to claim 1, wherein step 1, which involves estimating the fundamental frequency deviation using pre- and post-synchronization codes, comprises: Step 1.1: Calculate the center symbol angle of the preamble: The known symbol of the selected preamble is q, and the receiver receives the selected preamble symbol as q. ’ The selected preamble symbol number is Lq, and the selected preamble center symbol angle is... , where arg{} is the angle calculation function; Step 1.2: Calculate the center symbol angle of the postsynchronization code: The known symbol of the postsynchronization code is h, and the postsynchronization code symbol received by the receiver is h. ’ The number of symbols in the postsynchronization code is Lh, and the center symbol angle of the postsynchronization code is... ; Step 1.3: Calculate the frequency deviation: Ln is the number of symbols between the first symbol of the selected preamble and the last symbol of the postamble, Ls is the number of symbols between the center symbol of the selected preamble and the center symbol of the postamble, Ls = Ln - Lq / 2 - Lh / 2, Ts is the symbol period, and rem{} is the remainder function. Then the fundamental frequency deviation \Delta f = \frac {1} {2\pi Ts×Ls}\left \{{rem[Ah,2\pi ]-rem[Aq,2\pi ]} \right \} .

Images