Improved Gardner timing synchronization method suitable for burst transmission

Abstract

The invention discloses an improved Gardner timing synchronization method suitable for burst transmission. The method comprises the following steps: performing differential correlation, normalization and threshold judgment on received symbols, completing frame identifier detection, and obtaining frame identifier detection points; setting the frame identification detection point as an initial interpolation base point of a timing controller, and presetting a fractional interval according to the interpolation base point to complete initial phase injection; performing pre-filtering processing on the optimal sampling data after interpolation, and then calculating a Gardner timing error value; inputting the Gardner timing error value into a loop filter to generate a step length control signal; and the timing controller finally obtains and outputs optimal sampling data under the control of the step length control signal. According to the improved Gardner timing synchronization method suitable for burst transmission provided by the invention, an initial phase injection mechanism is introduced on the basis of a traditional Gardner timing synchronization algorithm, and a frame identification detection point is proposed to serve as an initial position of a Gardner algorithm interpolation point, so that the convergence speed of a system is remarkably accelerated.

Description

Technical Field

[0001] The present invention relates to the technical field of communication systems, and particularly to an improved Gardner timing synchronization method applicable to burst transmission. Background Art

[0002] In a communication system, a signal is affected by the Doppler effect during the channel transmission process. At the same time, due to a certain degree of mismatch between the clocks of the transmitter and the receiver, the signal received at the receiver may have frequency and phase offsets compared with the signal transmitted by the transmitter. When there are frequency offsets or phase offsets, the sampling clock cannot accurately sample the signal at the optimal sampling moment, resulting in a deviation between the sampled value and the ideal value. This deviation affects the subsequent symbol decision process, thereby reducing the signal quality and causing an increase in the bit error rate during the decoding process. Therefore, in order to obtain the optimal sampling moment of the symbol, a timing synchronization operation is required.

[0003] Regarding the timing synchronization of burst transmission systems, many studies have proposed different methods. For example, Wang Han and Wang Linnan used the traditional feedforward O&M algorithm to complete the timing synchronization of burst signals in "Research on Bit Timing Synchronization of Burst Signals" and analyzed the algorithm performance. Sun Jian et al. improved the O&M algorithm in "Improvement of Timing Synchronization Algorithm for Wireless Burst Communication" and performed sliding window processing on the data after segmented fast Fourier transform, which improved the convergence accuracy to a certain extent. However, these improved methods have not solved the problem of poor convergence accuracy of the traditional feedforward O&M, so their applicability in burst transmission scenarios is still limited. The Gardner timing error estimation algorithm is an estimation algorithm with a feedback structure. Its bit synchronization loop is simple and has a relatively high convergence accuracy, but its convergence speed is slow and it is not applicable to burst transmission systems. Summary of the Invention

[0004] In view of the above-mentioned defects of the prior art, the technical problems to be solved by the present invention are the problems such as low convergence accuracy of the existing communication system timing synchronization method and slow convergence speed based on the Gardner timing synchronization algorithm. The present invention provides an improved Gardner timing synchronization method applicable to burst transmission. On the basis of the traditional Gardner timing synchronization algorithm, an initial phase injection mechanism is introduced, and it is proposed to use the frame identification detection point as the initial position of the interpolation point of the Gardner algorithm, thereby significantly accelerating the convergence speed of the system.

[0005] To achieve the above object, the present invention provides an improved Gardner timing synchronization method applicable to burst transmission, including the following steps:

[0006] S1, perform differential correlation, normalization, and threshold decision on the received baseband signal to complete frame identification detection and obtain the frame identification detection point;

[0007] S2. Set the frame identification detection point as the starting interpolation base point of the timing controller, preset a fractional interval according to the interpolation base point, start the interpolation filter, complete the initial phase injection, and obtain the optimal sampled data after interpolation;

[0008] S3. Perform pre-filtering processing on the optimal sampled data after interpolation, and then send it to the Gardner timing error detector for timing error estimation to calculate the Gardner timing error value;

[0009] S4. Input the Gardner timing error value into the loop filter to generate a step size control signal;

[0010] S5. The timing controller determines the interpolation base point by underflow counting under the control of the step size control signal, generates a fractional interval signal, constructs a closed-loop feedback timing loop, and iterates cyclically to continuously adjust the interpolation base point, and finally obtains the optimal sampled data and outputs it.

[0011] Furthermore, perform differential correlation, normalization, and threshold decision on the received baseband signal to complete frame identification detection and obtain the frame identification detection point, which specifically includes the following steps:

[0012] The received baseband signal, that is, the baseband signal r after matched filtering k is divided into two paths. One path undergoes symbol delay to obtain r k-m , and one path undergoes conjugate processing to obtain Then perform complex multiplication on the two to complete the differential correlation operation;

[0013] Slide the result of the differential correlation operation point by point in each sampling point, perform correlation operation with the local frame identification differential template, and perform energy normalization to obtain the corresponding metric function result;

[0014] Set a normalization synchronization threshold. If the condition is met, it is considered that the frame identification is detected; at this time, the first sampling point of the differential correlation operation in step S1 is the frame identification detection point.

[0015] Furthermore, slide the result of the differential correlation operation point by point in each sampling point, perform correlation operation with the local frame identification differential template, and perform energy normalization to obtain the corresponding metric function result, and its corresponding calculation formula is as follows:

[0016]

[0017] where N is the frame synchronization identification length, μ is the current calculation moment, s is the constellation point mapped by the local frame identification sequence in the modulation mode, and L(μ) is the metric function of the sliding correlation.

[0018] Further, set a normalization synchronization threshold. If the following conditions are met,

[0019]

[0020] then it is considered that the frame identifier is detected. The physical meaning of the above relationship is: Continuously cache the values of 2M + 1 L(μ). If the central value is greater than the values of the remaining 2M points and the central value exceeds the set threshold.

[0021] Further, set the frame identifier detection point as the starting interpolation base point of the timing controller, and preset a fractional interval according to the interpolation base point. Then start the interpolation filter to complete the initial phase injection, which specifically includes:

[0022] In the frame identifier detection algorithm in step S1, capture the best frame identifier detection point. At this time, set the frame identifier detection point as the best sampling point of the first symbol. Then set the value of the phase decrementer to 0 and the corresponding fractional interval to 0. After that, start the interpolation filter in the Gardner loop to complete the initial phase injection.

[0023] Further, the timing controller is a phase decrementer, and in the Gardner timing synchronization loop, the starting interpolation base point m is determined by overflow counting under the action of the compensation control word w n k .

[0024] Further, in S3, perform pre-filtering processing on the interpolated best sampling data. Among them, the pre-filter is:

[0025]

[0026] Among them, H(f) is the transfer function of the system pulse shaping filter.

[0027] Further, send it to the Gardner timing error detector for timing error estimation. Among them, the Gardner timing error calculation formula is:

[0028]

[0029] Among them, e k is the estimated value of the timing error, y(n - 1) is the sampling value at the sampling moment of the (n - 1)-th symbol, y(n) is the sampling value at the sampling moment of the n-th symbol, is the intermediate sampling value between the sampling moments of the (n - 1)-th symbol and the n-th symbol, h(t) is the response of the pulse shaping filter of the communication system, and α is the roll-off coefficient of the pulse shaping filter.

[0030] Further, in S4, the Gardner timing error value is input into a loop filter to generate a step size control signal, and a digital ideal second-order active proportional-integral filter is used to implement the function of loop filtering.

[0031] Further, the transfer function expression corresponding to the digital ideal second-order active proportional-integral filter is: Among them, the corresponding Coe1 and Coe2 are the coefficients of the loop filter, and their specific calculation formulas are: Among them, w n represents the loop bandwidth; ξ represents the damping factor; K0 is the gain of the NCO; K d is the gain of the Gardner timing error detector; f s represents the operating frequency of the loop filter, that is, the frequency of the local sampling clock at the receiving end.

[0032] Technical Effects

[0033] An improved Gardner timing synchronization method applicable to burst transmission provided by the present invention, on the basis of the traditional Gardner timing synchronization algorithm, introduces an initial phase injection mechanism, and proposes a new structure for compensating absolute timing errors using frame identification detection points, thereby accelerating the convergence process of the Gardner timing recovery loop and significantly shortening the synchronization establishment time.

[0034] The following will further illustrate the concept, specific structure and technical effects generated by the present invention in conjunction with the drawings, so as to fully understand the purpose, features and effects of the present invention. Brief Description of the Drawings

[0035] Figure 1 is a schematic structural diagram of a traditional Gardner timing synchronization loop;

[0036] Figure 2 is a schematic structural diagram of a loop of an improved Gardner timing synchronization method applicable to burst transmission provided by an embodiment of the present invention;

[0037] Figure 3 is a frame identification detection flow chart of an improved Gardner timing synchronization method applicable to burst transmission provided by an embodiment of the present invention;

[0038] Figure 4 is a diagram of the update process of the phase value of the NCO register in an improved Gardner timing synchronization method applicable to burst transmission provided by an embodiment of the present invention;

[0039] Figure 5 is the phase discrimination characteristic corresponding to the Gardner timing error detector in an improved Gardner timing synchronization method applicable to burst transmission provided by an embodiment of the present invention;

[0040] Figure 6 It is a schematic diagram of the Gardner timing error detection algorithm in an improved Gardner timing synchronization method for burst transmission provided by an embodiment of the present invention;

[0041] Figure 7 It is a corresponding loop filter structure diagram in an improved Gardner timing synchronization method for burst transmission provided by an embodiment of the present invention;

[0042] Figure 8 It is a structural block diagram of a time - segmented parabolic interpolation filter in an improved Gardner timing synchronization method for burst transmission provided by an embodiment of the present invention;

[0043] Figure 9 It is a simulation verification result diagram based on the Matlab environment of an improved Gardner timing synchronization method for burst transmission provided by an embodiment of the present invention.

[0044] Figure 10 It is a comparison diagram of synchronization results between an improved Gardner timing synchronization method for burst transmission provided by an embodiment of the present invention and a traditional feed - forward O&M timing synchronization algorithm. Detailed implementation manners

[0045] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0046] In the following description, specific details such as specific internal programs and technologies are proposed for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of the present invention. However, those skilled in the art should clearly understand that the present invention can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well - known systems, devices, circuits and methods are omitted to avoid unnecessary details from interfering with the description of the present invention.

[0047] As Figure 1 shown, the existing traditional Gardner timing synchronization loop is a feedback loop, in which the timing error detector uses the Gardner algorithm to calculate the timing error value at each symbol point and outputs it to the loop filter for smoothing processing. The timing controller generates interpolation base points and fractional intervals using the control word output by the loop filter, and finally completes timing correction through the interpolation filter; since it is a feedback loop, it takes a long time to fully converge.

[0048] As Figure 2As shown in the figure, the principle of an improved Gardner timing synchronization method applicable to burst transmission provided by the present invention is as follows: On the basis of the traditional Gardner timing synchronization loop, an initial phase injection mechanism is introduced. First, frame identification detection is completed through differential correlation and threshold decision to determine the position of the frame identification detection point, and then the timing controller in the Gardner loop is initialized using the frame identification detection point to complete the initial convergence of the loop.

[0049] The present invention provides an improved Gardner timing synchronization method applicable to burst transmission, and the specific implementation steps are as follows:

[0050] S1, as Figure 3 shown in the figure, perform differential correlation, normalization, and threshold decision on the received baseband signal to complete frame identification detection;

[0051] Step S1 specifically includes the following steps:

[0052] S11, perform a delay conjugate correlation operation on the baseband signal input at the receiving end. The baseband signal r k is divided into two paths. One path is symbol-delayed to obtain r k-m , and the other path is conjugated to obtain Then, perform a complex multiplication operation on the two, and its corresponding calculation formula is as follows:

[0053]

[0054] where Z k is the delay conjugate correlation result, and m is the number of sampling points in the symbol interval.

[0055] S12, slide the result of the delay conjugate correlation operation point by point in sampling, perform a correlation operation with the local frame identification differential template s j , and perform energy normalization to obtain the corresponding metric function result. Its corresponding calculation formula is as follows:

[0056]

[0057] where L is the frame synchronization identification length, μ is the current calculation moment, s is the constellation point mapped by the local frame identification sequence in the modulation mode, and L(μ) is the metric function of the sliding correlation.

[0058] S13: Set the normalization synchronization threshold. If the following conditions are met

[0059]

[0060] Then it is considered that the frame identifier is detected. The physical meaning of the above relationship is: Continuously buffer the values of 2M + 1 L(μ). If the central value is greater than the values of the other 2M points and the central value exceeds the set threshold.

[0061] When the frame identifier is detected, the first sampling point for the differential correlation operation in step S1 is the frame identifier detection point.

[0062] In this embodiment, the decision threshold η0 = 0.45 is set, and M = 5 is taken. If the relationship is satisfied:

[0063]

[0064] Its corresponding physical relationship is: Continuously buffer the values of 11 L(μ). If the central value is greater than the values of the other 10 points and the central value exceeds the set threshold of 0.45, then it is considered that the frame identifier is captured.

[0065] S2. Set the frame identifier detection point obtained in step S1 as the starting interpolation base point of the timing controller, preset the fractional interval according to the interpolation base point, start the interpolation filter, and complete the initial phase injection.

[0066] Specifically, the timing controller is a phase decrementer, that is, an NCO. In the Gardner timing synchronization loop, the interpolation base point m is determined by overflow counting under the action of the compensation control word w n The function of the phase decrementer can be expressed by the following formula: k

[0067]

[0068] Among them, η m represents the value of the phase decrementer when the m-th symbol sampling moment arrives, η(m + 1) represents the value of the phase decrementer when the (m + 1)-th symbol sampling moment arrives, and mod1 represents the modulo 1 operation.

[0069] Such as Figure 4 intuitively shows the update process of the NCO register phase value. From it, the fractional interval μ can be derived using the similar triangle rule:

[0070]

[0071] Among them, μ k represents the fractional interval, m k represents the interpolation base point, and T s represents the sampling clock period. After deforming it by one bit, the expression of the fractional interval μ k can be obtained:

[0072]

[0073] Among them, w(m k ) represents the compensation control word output by the loop filter. When the timing synchronization loop converges, w(m) is approximately equal to a constant. At this time, the NCO performs a zero-crossing detection once every 1 / w(m) working cycles, that is Thus, the relationship expression between the step control word, the interpolation adjustment period, and the sampling clock period can be derived:

[0074] Specifically, in the frame identification detection algorithm in step S1, this method can capture the best frame identification detection point. At this time, set the frame identification detection point as the best sampling point of the first symbol, and then set the initial value η(0) of the phase decrementer to 0 and the corresponding initial fractional interval μ0 to 0 to complete the initial phase injection. This is an example of an initial phase injection.

[0075] S3. Perform pre-filtering processing on the interpolated best sampling data, and then send it to the Gardner timing error detector for timing error estimation, specifically as Figure 5 shown.

[0076] Figure 5 In the figure is the phase discrimination characteristic of the Gardner timing error detector. The mean value of the loop timing error is only related to the roll-off coefficient of the shaping filter, that is, within the bandwidth of (1-α) / 2T′~(1+α) / 2T, the existence of the remaining frequency band components acts as the so-called self-noise. Specifically, the pre-filter corresponding to step S3 is:

[0077]

[0078] Among them, H(f) is the transfer function of the system pulse shaping filter.

[0079] Specifically, the Gardner timing error calculation formula corresponding to step S3 is:

[0080]

[0081] Among them, e n is the estimated value of the timing error, y(n-1) is the sampling value at the sampling moment of the (n-1)th symbol, y(n) is the sampling value at the sampling moment of the nth symbol, is the intermediate sampling value between the sampling moments of the (n-1)th symbol and the nth symbol, h(t) is the response of the communication system pulse shaping filter, and α is the roll-off coefficient of the pulse shaping filter. As Figure 6As shown, in a), the middle sampling value is greater than zero, indicating that the sampling clock is ahead, and interpolation needs to be adjusted backward; in b), the middle sampling value is zero, and at this time the timing error output value is zero, which is the optimal sampling point of the loop: in c), the middle sampling value is less than zero, indicating that the sampling clock is lagging, and interpolation needs to be adjusted forward.

[0082] S4. Input the Gardner timing error value calculated above into the loop filter to generate a step control signal.

[0083] Specifically, for the loop filter in step S4, its working principle is as Figure 7 shown. In this embodiment, a digital ideal second-order active proportional-integral filter is used to implement the function of loop filtering, and the corresponding transfer function expression is: where Coe1 and Coe2 corresponding to it are the coefficients of the loop filter.

[0084] Their specific calculation formulas are: where w n represents the loop bandwidth; ξ represents the damping factor; K0 is the gain of the NCO; K d is the gain of the Gardner timing error detector; f s represents the operating frequency of the loop filter, that is, the frequency of the local sampling clock at the receiving end. According to the above, the recursive equation of the loop filter can be obtained as:

[0085] w(n) = w(n - 1) + Coe1·[e(n) - e(n - 1)] + Coe2·e(n)

[0086] In this embodiment, the total loop gain K0·K d = 1 is set, the damping factor ξ = 0.707 is set, and the loop bandwidth calculation formula is:

[0087]

[0088] where B represents the noise bandwidth of the loop filter, generally taking a value of 0.001 of the symbol rate. Substituting all such parameters into the calculation formulas of Coe1 and Coe2, the specific recursive equation can be obtained:

[0089] w(n) = w(n - 1) + 2.7×10 -3 ×[e(n) - e(n - 1)] + 3.56×10 -6 ×e(n)

[0090] S5. Under the control of the step control signal, the timing controller determines the interpolation base point by the underflow counting method, generates a fractional interval signal, constructs a closed-loop feedback timing loop, and iterates cyclically to continuously adjust the interpolation base point, and finally obtains the data at the optimal sampling moment and outputs it.

[0091] Specifically, in step S5, the input signal is interpolated using a piecewise parabolic interpolation filter with α = 0.5 to obtain the data at the optimal sampling moment; the structural block diagram of the interpolation filter is as Figure 8 shown. The order N of the interpolation filter is 4, and the corresponding interpolation coefficients are as follows:

[0092]

[0093] Specifically, the calculation formula for the interpolation point is:

[0094] y(kT i ) = C -2 (μ k )x(m k + 2) + C -1 (μ k )x(m k + 1) + C0(μ k )x(m k ) + C1(μ k )x(m k - 1)

[0095] To verify the effectiveness of the timing synchronization method of the present invention, a specific example will be given below for simulation testing:

[0096] 1. Modulated with a frame header of , and the data is in the 32APSK modulation format.

[0097] 2. E s / N0 is 50 dB, and the roll-off factor is 0.2.

[0098] Simulation content and result analysis:

[0099] As Figure 2 is the system structure diagram of the present invention. First, step S1 is performed to complete the detection of the frame identification point; the corresponding relevant results are as Figure 9 shown in a. After the frame identification detection, there is an obvious peak point in the figure. Then, the initial phase injection is completed through step S2. For more intuitive results during simulation, η(0) of the timing controller is set to 0.2, and the corresponding μ0 is set to 0.4. The corresponding interpolation base point should be shifted two sampling points forward relative to the frame identification detection point. Subsequently, step S3 is completed to obtain the estimated timing error. The loop filter generates a step control signal w(n) based on the estimated timing error, and the timing controller uses the w(n) output by the loop filter to generate the interpolation base point and the fractional interval μ k ; then, the timing correction is completed through the interpolation filter. As shown in Figure 9 , Figure 9 b is the error output by the corresponding Gardner timing error detector.Figure 9 c is the output corresponding to the loop filter, Figure 9 d is the fractional interval output by the corresponding timing controller. Finally, the data output by the interpolation filter is decimated at twice the symbol rate to obtain the symbols after timing synchronization, as Figure 10 shown in Fig. f.

[0100] Figure 10 shows the comparison of the convergence speed and convergence accuracy between an improved Gardner timing synchronization method applicable to burst transmission proposed by the present invention and the traditional feed-forward O&M timing synchronization algorithm and the traditional feedback Gardner timing synchronization algorithm. In the figure, one Data Block represents 256 symbols. Further, the normalized mean square error between the constellation points recovered by the algorithm and the standard constellation points and the constellation point situation of the output are used as the measurement criteria for the timing performance. From Figure 10 Figs. 10a and 10d, it can be seen that the traditional feed-forward O&M timing synchronization algorithm has no convergence time, but the convergence accuracy is poor, and there is a large loss when the system modulation order is high. Figure 10 Figs. 10b and 10e are the synchronization situations of the traditional feedback Gardner timing synchronization algorithm. Although the convergence accuracy is high, it requires a certain convergence time (>2000 symbols). Figure 10 Figs. 10c and 10f are the timing synchronization situations of the structure proposed by the present invention, which takes into account both the convergence speed and the convergence accuracy. The loop can directly converge initially and complete the full convergence of the feedback loop within 512 symbols. This result proves that the improved algorithm of the present invention effectively solves the problems of poor convergence accuracy of the traditional feed-forward O&M timing synchronization algorithm and slow convergence speed of the traditional feedback Gardner algorithm, is applicable to burst transmission systems, and has the ability to process high-order modulation signals under low roll-off factor conditions.

[0101] The preferred specific embodiments of the present invention have been described in detail above. It should be understood that those of ordinary skill in the art can make many modifications and variations based on the concept of the present invention without creative work. Therefore, all technical solutions that can be obtained by those skilled in the art in the technical field of the present invention based on the concept of the present invention through logical analysis, reasoning or limited experiments on the basis of the prior art should be within the protection scope determined by the claims.

Claims

1. An improved Gardner timing synchronization method suitable for burst transmission, characterized in that: The following steps are involved: S1, perform differential correlation, normalization, and threshold judgment on the received baseband signal, complete frame mark detection, and obtain frame mark detection points; S2, setting the frame identification detection point as the starting interpolation base point of the timing controller, and according to the preset fractional interval of the interpolation base point, starting the interpolation filter, completing the initial phase injection, and obtaining the best sampling data after interpolation; S3, pre-filtering the interpolated optimal sampling data, and then sending it to the Gardner timing error detector for timing error estimation, and calculating the Gardner timing error value; S4, inputting the Gardner timing error value into a loop filter to generate a step control signal; S5, the timing controller determines the interpolation base point by underflow counting under the control of the step control signal, generates a fractional interval signal, builds a closed-loop feedback timing loop, iterates to continuously adjust the interpolation base point, and finally obtains the best sampling data and outputs it.

2. An improved Gardner timing synchronization method suitable for burst transmission as claimed in claim 1, characterized in that: The received baseband signal is differentially correlated, normalized, and thresholded to complete frame identification detection and obtain a frame identification detection point, which specifically includes the following steps: The received baseband signal, that is, the baseband signal after matched filtering r k Divide into two paths, one path is symbol delayed to obtain r k-m , and conjugate treatment is performed all the way to obtain Then the two are multiplied by complex numbers to complete the differential correlation operation; Slide the result of the differential correlation operation one sampling point at a time, perform correlation operation with the local frame identification differential template, and perform energy normalization to obtain the corresponding metric function result; A normalized synchronization threshold is set. If the condition is met, it is considered that the frame mark is detected. At this time, the first sampling point of the differential correlation operation in step S1 is the frame mark detection point.

3. An improved Gardner timing synchronization method suitable for burst transmission as claimed in claim 2, characterized in that: The result of the differential correlation operation is slid sampling point by sampling point, and the correlation operation is performed with the local frame identification differential template, and the energy is normalized to obtain the corresponding metric function result, and the corresponding calculation formula is as follows: Where N is the frame synchronization identifier length, μ is the current calculation time, and s is the local frame identifier sequence. The constellation point mapped by the modulation mode, L(μ) is the sliding related metric function.

4. An improved Gardner timing synchronization method suitable for burst transmission as claimed in claim 2, characterized in that: Set the normalized synchronization threshold. If the following conditions are met, It is considered that the frame identifier is detected. The physical meaning of the above relationship is: 2M+1 L(μ) values ​​are cached continuously. If the central value is greater than the values ​​of the remaining 2M points, and the central value exceeds the set threshold.

5. The improved Gardner timing synchronization method suitable for burst transmission as claimed in claim 1, characterized in that: The frame identification detection point is set as the starting interpolation base point of the timing controller, and the fractional interval is preset according to the interpolation base point, and the interpolation filter is started to complete the initial phase injection, which specifically includes: The frame marker detection algorithm in step S1 captures the best frame marker detection point. At this time, the frame marker detection point is set as the best sampling point of the first symbol. Then, the value of the phase decrementer is set to 0 and the corresponding fractional interval is set to 0. Then, the interpolation filter in the Gardner loop is started to complete the initial phase injection.

6. An improved Gardner timing synchronization method suitable for burst transmission as claimed in claim 5, characterized in that: The timing controller is a phase decrementer that compensates the control word w in the Gardner timing synchronization loop. n The starting interpolation base point m is determined by overflow counting under the action of k .

7. An improved Gardner timing synchronization method suitable for burst transmission as claimed in claim 1, characterized in that: S3, pre-filtering the interpolated optimal sampling data, wherein the pre-filter is: Where H(f) is the system pulse shaping filter transfer function.

8. An improved Gardner timing synchronization method suitable for burst transmission as claimed in claim 7, characterized in that: The signal is sent to the Gardner timing error detector for timing error estimation, where the Gardner timing error calculation formula is: Among them, e k is the estimated value of the timing error, y(n-1) is the sampling value at the sampling time of the n-1th symbol, y(n) is the sampling value at the sampling time of the nth symbol, is the intermediate sampling value between the n-1th symbol sampling time and the nth symbol sampling time, h(t) is the response of the pulse shaping filter of the communication system, and α is the roll-off coefficient of the pulse shaping filter.

9. The improved Gardner timing synchronization method suitable for burst transmission as claimed in claim 1, characterized in that: S4, inputting the Gardner timing error value into the loop filter, generating a step control signal, and using a digital ideal second-order active proportional integral filter to implement the loop filtering function.

10. An improved Gardner timing synchronization method suitable for burst transmission as claimed in claim 9, characterized in that: The corresponding transfer function expression of the digital ideal second-order active proportional integral filter is: Among them, the corresponding Coe1 and Coe2 are the coefficients of the loop filter, and their specific calculation formulas are: Among them, w n represents the loop bandwidth; ξ represents the damping factor; K0 is the gain of NCO; K d is the gain of the Gardner timing error detector; f s Indicates the operating frequency of the loop filter, that is, the frequency of the local sampling clock at the receiving end.

Images