CN116388755A - Improved timing synchronization method for initial phase injection - Google Patents

Abstract

The invention provides an improved timing synchronization method for initial phase injection, which comprises the following steps of S1, performing square filtering timing error estimation on an input signal to obtain a square filtering timing error value; s2, taking the square filtering timing error value as an initial position of an interpolation point of a Gardner timing loop, and carrying out Gardner timing error estimation on the optimal sampling data obtained after interpolation; s3, inputting the timing error value into the loop filter, the gain control filter and the selection controller at the same time, adjusting the loop coefficient through the selection controller, and multiplying the sum of the outputs of the loop filter and the gain control filter with the loop coefficient to be used as a control word of the numerical control oscillator; s4, generating an updating value of a fractional interval by the numerical control oscillator in an overflow counting mode, changing the position of an interpolation base point, forming a timing synchronization loop of closed-loop feedback, and continuously adjusting the interpolation base point by loop iteration to obtain optimal sampling data; s5, carrying out downsampling filtering on the optimal sampling data to output the optimal sampling data.

Description

Improved timing synchronization method for initial phase injection

Technical Field

The application relates to the technical field of communication, in particular to an improved timing synchronization method for initial phase injection.

Background

In a communication system, signals are affected by doppler effect in the channel transmission process, and in addition, clocks of a transmitting end and a receiving end are not completely matched, so that signals received by the receiving end and signals sent by the transmitting end have certain frequency and phase offset. When the frequency offset or the phase offset exists, the sampling clock cannot sample the signal at the optimal sampling time, and the obtained sampling value deviates from the ideal value, and the deviation can influence the symbol judgment process of subsequent demodulation, reduce the signal quality and cause the increase of the bit error rate in the decoding process, so that the bit synchronization processing is needed, and the frequency and the phase offset of the sampling clock are compensated and corrected.

The initial bit synchronization method is mainly an insert pilot method, which requires inserting a piece of pilot information into a data frame, and the receiver performs bit synchronization according to the piece of pilot information. The pilot frequency information occupies a large amount of transmission energy, increases frequency spectrum resources, influences the throughput rate of signal transmission, and is less adopted in an actual communication system.

The bit synchronization method commonly used at present mainly comprises a square filtering timing estimation algorithm and a Gardner timing error estimation algorithm. The square filtering timing estimation algorithm belongs to a feedforward structure estimation algorithm, and the method firstly carries out square nonlinear transformation on a received signal, and then extracts a frequency component at 1/T through digital filtering, so as to obtain an estimated value of timing deviation. The square filtering timing estimation algorithm has high convergence speed, but has high requirement on the sampling rate of the ADC, at least four times of oversampling is needed, and the algorithm has poor tracking performance and cannot resist larger timing deviation. The Gardner timing error estimation algorithm is a feedback structure estimation algorithm that requires only two sample points for each symbol timing error value calculation. The Gardner timing error estimation algorithm is widely used because of its simple structure of the bit synchronization loop and the synchronization performance is not affected by the carrier phase interference, but its convergence speed is slow, which is not suitable for burst transmission.

Disclosure of Invention

The invention provides an improved timing synchronization method of initial phase injection, which comprises the following steps:

s1: square filtering timing error estimation is carried out on an input signal, a square filtering timing error value is obtained, and clock initial phase capturing is completed;

s2: taking the square filtering timing error value as an initial position of an interpolation point of a Gardner timing loop, and carrying out Gardner timing error estimation on the optimal sampling data obtained after interpolation;

s3: the Gardner timing error value obtained through calculation is input into a loop filter, a gain control filter and a selection controller at the same time, loop coefficients are adjusted through the selection controller, and the sum of the outputs of the loop filter and the gain control filter is multiplied by the loop coefficients to be used as a control word of the numerical control oscillator;

s4: the numerical control oscillator generates an updating value of a fractional interval in an overflow counting mode, changes the position of an interpolation base point to form a timing synchronization loop of closed-loop feedback, and then carries out cyclic iteration to continuously adjust the interpolation base point to obtain optimal sampling data;

s5: and carrying out downsampling filtering on the optimal sampling data, and outputting the optimal sampling data after downsampling filtering.

Further, in step 1, the method further comprises the steps of:

s11: initializing a control word of a first numerical control oscillator and simultaneously acquiring input data;

s12: the input data is sent to an interpolation filter, the data is output under the control of a first numerical control oscillator, and when the number of the output data reaches a preset value, the estimation of a clock initial phase is carried out;

s13: and calculating to obtain a clock initial phase estimation by a square filtering algorithm, and finishing the initialization of the second digital controlled oscillator according to the clock initial phase estimation.

Further, in step S13, the square filtering algorithm performs square nonlinear transformation on the received signal, and then extracts the frequency component at 1/T through digital filtering, thereby obtaining an estimated value of the timing deviation

The calculation formula is as follows:

wherein r (k) represents data output by the interpolation filter under the control of the first numerical control oscillator, k represents sampling points, N represents sampling points in one symbol period, L represents symbol length, generally, L can meet the requirement by taking 64, and arg represents angle taking operation;

and when the number of the output data reaches LN, estimating the initial phase of the clock.

Further, in step S2, the method further includes the steps of:

s21: the input data is calculated to obtain an interpolation sequence of a loop through an interpolation filter under the control of a second numerical control oscillator;

s22: and performing timing error estimation operation on the interpolation sequence according to the Gardner algorithm to obtain a timing error signal.

Further, in step S3, the method further includes the steps of:

s31: inputting the timing error into a loop filter to filter out high frequency components, and simultaneously inputting the timing error into a gain control filter;

s32: the selection controller judges the loop state by calculating the error average value of the timing error, thereby adjusting the loop coefficient;

s33: the result of multiplying the sum of the loop filter and the gain control filter output by the loop coefficient is used as the control word for the second numerically controlled oscillator.

Further, in step S4, the method further includes the steps of:

s41: the second numerically controlled oscillator determines interpolation base points and fractional intervals required by the interpolation filter in an overflow counting mode;

s42: the interpolation filter completes a new round of interpolation calculation by utilizing the interpolation base points and the fractional intervals, thereby forming a closed-loop feedback loop, and continuously adjusting the positions of the interpolation base points through cyclic iteration to obtain optimal sampling data.

The beneficial effects of this application are:

the method combines the advantages of the square filtering timing estimation algorithm and the Gardner timing error estimation algorithm, and takes the square filtering timing error value as the initial position of the interpolation point of the Gardner algorithm, so that the Gardner timing recovery loop is quickly locked, and the time for establishing timing synchronization is shortened.

The present invention adds a gain control filter, a selection controller, and loop coefficients to a conventional Gardner timing recovery loop. The sum of the loop filter and the output of the gain control filter is multiplied by the loop coefficient to be used as a control word of the numerical control oscillator, wherein the selection controller is used for judging the loop state so as to adjust the loop coefficient. The improved loop structure can effectively reduce timing jitter errors caused by self-noise of the system in the synchronization process.

Drawings

The advantages of the foregoing and/or additional aspects of the present application will become apparent and readily appreciated from the description of the embodiments, taken in conjunction with the accompanying drawings, wherein:

FIG. 1 is a flow chart of an improved method for timing synchronization for initial phase injection provided by an embodiment of the present invention;

FIG. 2 is a schematic structural diagram of an improved timing synchronization device for initial phase injection according to an embodiment of the present invention;

fig. 3 is a diagram showing an update procedure of an NCO register phase value in an improved timing synchronization method for initial phase injection according to an embodiment of the present invention.

Fig. 4 is a schematic diagram of a square filtering timing error estimation algorithm in an improved timing synchronization method for initial phase injection according to an embodiment of the present invention.

Fig. 5 is a schematic diagram of a Gardner timing error detection algorithm in an improved timing synchronization method for initial phase injection according to one embodiment of the present invention.

Fig. 6 is a schematic diagram of the operation of a gain control filter in an improved timing synchronization method for initial phase injection according to an embodiment of the present invention.

Fig. 7 is a schematic diagram of the operation of the selection controller in the improved timing synchronization method for initial phase injection according to the embodiment of the present invention.

Fig. 8 is a block diagram of a cubic interpolation filter in an improved timing synchronization method for initial phase injection according to an embodiment of the present invention.

Detailed Description

In order that the above-recited objects, features and advantages of the present application will be more clearly understood, a more particular description of the application will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, without conflict, the embodiments of the present application and features of the embodiments may be combined with each other.

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application, however, the present application may be practiced otherwise than as described herein, and thus the scope of the present application is not limited to the specific embodiments disclosed below.

As shown in fig. 1, an improved timing synchronization method for initial phase injection, the timing synchronization method comprises the following steps:

s1: square filtering timing error estimation is carried out on an input signal, a square filtering timing error value is obtained, and clock initial phase capturing is completed;

specifically, in step S1, the method further includes:

s11: initializing a first numerical control oscillator NCO1, thereby initializing an initial value of an NCO1 register, initializing a control word, and simultaneously sending input data into a data buffer 1;

s12: closing a switch to k1, taking out data from the data buffer 1, entering the interpolation filter 1, continuously storing the data output by the interpolation filter 1 into the data buffer 2 under the control of the first numerical control oscillator NCO1, and estimating the initial phase of the clock when the number of stored data reaches LN;

s13: calculating to obtain the estimated value of the initial phase of the clock by a square filtering algorithm

And completes the numerical control oscillator NInitialization of CO 2.

Specifically, the first numerically controlled oscillator NCO1 and the numerically controlled oscillator NCO2 are both phase decrementers, and the interpolation base point m is determined by means of overflow counting under the action of the step control word ω (m) _k Wherein the step control word of the first numerically controlled oscillator NCO1 is a predetermined constant; the step control word ω (m) of the numerically controlled oscillator NCO2 is provided by the loop filter. The function of the phase decrementer can be formulated:

η (m) represents the value of the NCO register when the mth sampling instant arrives, η (m+1) represents the value of the NCO register when the mth+1th sampling instant arrives, mod1 represents a modulo 1 operation.

Fig. 3 shows intuitively the update procedure of the phase values of the NCO1, NCO2 registers, from which it can be deduced using the rule of similar triangles:

wherein mu _k Represents fractional intervals, m _k Representing the interpolation base point, T _s Representing the sampling clock period.

After being shifted and deformed, the product can obtain a fraction interval mu _k Is represented by the expression:

wherein ω (m) _k ) A step control word representing the loop filter output.

In the present embodiment, since the input data is 4 times oversampled data, the control word ω of the numerically controlled oscillator NCO1 is set to _NCO1 (m) initial value η of register set to constant 1, NCO1 _NCO (1) Set to 1.

Specifically, in step S13, the square filtering algorithm,as shown in figure 4, the method firstly carries out square nonlinear transformation on the received signal, then extracts the frequency component at 1/T through digital filtering, thereby obtaining the estimated value of the timing deviation

The calculation formula is as follows:

where r (k) represents data output by the interpolation filter under the control of the first numerically controlled oscillator, k represents sampling points, N represents sampling points in one symbol period, L represents symbol length, generally, taking 64 can satisfy the requirement, and arg represents angular operation. In the present embodiment, the first 256 input symbols are used to calculate the timing offset estimate

I.e., N is 4 and L is 64.

Obtaining timing deviation estimated value

The numerically controlled oscillator NCO2 can then be initialized similar to NCO1 but with NCO2 having control word ω _NC (m) is not constant, and its overflow period is dynamically variable. Since N is taken to be 4, an initial value omega of an NCO2 control word is set _NCO2 (1) An initial value η of the NCO2 register of 0.5 _NC (1) Is->

After the initialization setting of NCO2 is completed, step S2 is entered.

S2: taking the square filtering timing error value as an initial position of an interpolation point of a Gardner timing loop, and carrying out Gardner timing error estimation on the optimal sampling data obtained after interpolation;

specifically, in step S2, the method further includes:

s21: closing a switch to k2, and calculating the data in the data buffer 1 through an interpolation filter 2 under the control of a numerical control oscillator NCO2 to obtain an interpolation sequence of a loop;

s22: the timing error detector carries out timing error estimation operation on the interpolation sequence according to the Gardner algorithm to obtain a timing error signal e (n).

Specifically, in step S22, timing error estimation is performed by using the Gardner algorithm, where the expression of the conventional Gardner timing error detection algorithm is:

wherein e (n) is the calculated timing error value, y _I (n)、y _Q (n) symbol sample values respectively representing the n-th in-phase signal and the quadrature signal, y _I (n-1)、y _Q (n-1) represents symbol sample values of the n-1 th in-phase signal and the quadrature signal, respectively,

representing the mid-sample values of the in-phase signal, the n-1 th and the n-th sample instants of the quadrature signal, respectively. As shown in fig. 5 a), in a), the intermediate sampling value is zero, and the timing error output value is zero at this time, which is the optimal sampling point of the loop; b) Wherein, the middle sampling value is larger than zero, which indicates that the sampling clock is advanced, and interpolation needs to be adjusted backwards; c) In which the intermediate sample value is less than zero, indicating that the sample clock lags and the interpolation needs to be adjusted forward.

S3: the Gardner timing error value obtained through calculation is input into a loop filter, a gain control filter and a selection controller at the same time, loop coefficients are adjusted through the selection controller, and the sum of the outputs of the loop filter and the gain control filter is multiplied by the loop coefficients to be used as a control word of the numerical control oscillator;

specifically, in step S3, the method further includes:

s31: inputting the detected timing error e (n) into a loop filter to filter out a high frequency component, and simultaneously inputting the timing error e (n) into a gain control filter;

s32: the selection controller judges the loop state by calculating the error mean value, thereby adjusting the loop coefficient;

s33: the result of multiplying the sum of the loop filter and the gain control filter output by the loop coefficient is used as a control word omega (m) of the numerically controlled oscillator NCO2, and the value of the NCO register 2 is updated.

Further, the loop filter in step S31 is mainly used for filtering out high-frequency interference components in the timing error signal, and in this embodiment, a second-order active proportional-integral filter is used to implement the function of loop filtering, and the corresponding transfer function expression is as follows:

wherein coe and coe are coefficients of the loop filter, and the specific calculation formula is +.>

ω _n Represents loop bandwidth, ζ represents damping factor, K _m Is the gain of NCO, K _d Is the phase discrimination gain of the timing error detector, f _s Indicating the operating frequency of the loop filter, i.e. the frequency of the local sampling clock at the receiving end. The recursive equation for the loop filter is:

ω(n)＝ω(n-1)+coe1·[e(n)-e(n-1)]+coe2·e(n)

specifically, the gain control filter described in step S31, whose operation schematic diagram is shown in fig. 6, sets an error threshold u in the filter _ref By detecting whether the timing error e (n) exceeds the threshold, if the continuous input exceeds a certain number m, an error value e' (n) =sign (e (n)) (u) is output _ref ) Otherwise, the error value e' (n) =0 is output. In the present embodiment, m is set to 3, u through parameter simulation _ref The gain control filter performs best when set to 0.2。

Further, the operation schematic diagram of the selection controller in step S32 is shown in fig. 7, and the controller determines the loop state by calculating the average value of the error values, and the timing synchronization loop enters the tracking state when the average value approaches zero, otherwise, the controller is in the capturing state. Setting a loop coefficient to 1 when the loop is in a capture state; when in tracking state, the loop coefficient should take a value between 0 and 1, and the timing jitter and the synchronization time are comprehensively analyzed through simulation experiments, and finally the loop coefficient value is selected. In this embodiment, when the loop is in the tracking state, the timing jitter and the synchronization time are comprehensively analyzed through a simulation experiment, and finally the loop coefficient is set to 0.1.

S4: the numerical control oscillator generates an updating value of a fractional interval in an overflow counting mode, changes the position of an interpolation base point to form a timing synchronization loop of closed-loop feedback, and then carries out cyclic iteration to continuously adjust the interpolation base point to obtain optimal sampling data;

specifically, in step S4, the method further includes:

s41: the numerically controlled oscillator NCO2 determines the interpolation base point m required for the interpolation filter 2 by means of overflow counting under control of ω (m) _k And fractional spacing mu _k ；

S42: the interpolation filter 2 uses m _k Sum mu _k And finishing a new round of interpolation calculation, thereby forming a closed-loop feedback loop, and continuously adjusting the interpolation base point position through loop iteration to obtain the optimal sampling data.

Specifically, the interpolation filter described in steps S1, S2 generally employs a polynomial interpolation filter whose basic principle is to approximate the interpolation point y (kT) by using a lagrangian polynomial function of order N _i ) Obtaining a Lagrange interpolation formula expression:

wherein C is _i (t) represents the Lagrangian coefficient expressed in the following manner

Since the number N of interpolator tap coefficients involved in interpolation point operation must be even, I can be set ₁ ＝-N/2，I ₂ =n/2-1. Because only the interpolation filter unit impulse response is required to be within (mu) _k +i)T _s Sampling value h of time _I [(μ _k +i)T _s ]So let t _μ ＝(μ _k +i)T _s The expression for the lagrangian interpolation filter can be obtained:

h _I [(μ _k +i)T _s ]＝C _i (μ _k )

wherein, the liquid crystal display device comprises a liquid crystal display device,

in this example, the cubic interpolation filter with the Farrow structure is shown in fig. 8, and the order of the interpolation filter is n=4, I ₁ ＝-2，I ₂ =1, cubic interpolation coefficient of

The calculation formula of the interpolation point is y (kT _i )＝C _-2 (μ _k )x(m _k +2)+C _-1 (μ _k )x(m _k +1)+C ₀ (μ _k )x(m _k )+C ₁ (μ _k )x(m _k -1)。

S5: and carrying out downsampling filtering on the optimal sampling data, and outputting the optimal sampling data after downsampling filtering.

The invention designs an improved timing synchronization method for initial phase injection, which uses a method of combining a square filtering timing estimation algorithm and a Gardner timing error estimation algorithm to perform timing synchronization on signals. Firstly, a signal finishes clock initial phase capture through square filtering error estimation, and then takes a square filtering timing error value as an initial position of a Gardner algorithm interpolation point, so that a Gardner timing recovery loop is quickly locked, and the time for establishing timing synchronization is shortened; and a gain control filter, a selection controller and loop coefficients are added in a traditional Gardner timing recovery loop, so that the timing jitter error caused by self-noise of the system in the synchronization process is effectively reduced. Compared with the traditional Gardner timing recovery algorithm, the method can effectively shorten the synchronization establishment time and reduce the timing jitter error.

The steps in the present application may be sequentially adjusted, combined, and pruned according to actual requirements.

Although the present application is disclosed in detail with reference to the accompanying drawings, it is to be understood that such descriptions are merely illustrative and are not intended to limit the application of the present application. The scope of the present application is defined by the appended claims and may include various modifications, alterations, and equivalents to the invention without departing from the scope and spirit of the application.

Claims (6)

1. An improved timing synchronization method for initial phase injection, characterized in that the timing synchronization method for initial phase injection comprises the following steps:

s1: square filtering timing error estimation is carried out on an input signal, a square filtering timing error value is obtained, and clock initial phase capturing is completed;

s2: taking the square filtering timing error value as an initial position of an interpolation point of a Gardner timing loop, and carrying out Gardner timing error estimation on the optimal sampling data obtained after interpolation;

s3: the Gardner timing error value obtained through calculation is input into a loop filter, a gain control filter and a selection controller at the same time, loop coefficients are adjusted through the selection controller, and the sum of the outputs of the loop filter and the gain control filter is multiplied by the loop coefficients to be used as a control word of the numerical control oscillator;

s4: the numerical control oscillator generates an updating value of a fractional interval in an overflow counting mode, changes the position of an interpolation base point to form a timing synchronization loop of closed-loop feedback, and then carries out cyclic iteration to continuously adjust the interpolation base point to obtain optimal sampling data;

s5: and carrying out downsampling filtering on the optimal sampling data, and outputting the optimal sampling data after downsampling filtering.

2. The timing synchronization method of initial phase injection according to claim 1, further comprising the steps of, in step 1:

s11: initializing a control word of a first numerical control oscillator and simultaneously acquiring input data;

s12: the input data is sent to an interpolation filter, the data is output under the control of a first numerical control oscillator, and when the number of the output data reaches a preset value, the estimation of a clock initial phase is carried out;

s13: and calculating to obtain a clock initial phase estimation by a square filtering algorithm, and finishing the initialization of the second digital controlled oscillator according to the clock initial phase estimation.

3. The method of timing synchronization for initial phase injection according to claim 1, wherein in step S13, the square filtering algorithm performs square nonlinear transformation on the received signal, and then extracts frequency components at 1/T through digital filtering, thereby obtaining an estimated value of timing deviation

The calculation formula is as follows:

wherein r (k) represents data output by the interpolation filter under the control of the first numerical control oscillator, k represents sampling points, N represents sampling points in one symbol period, L represents symbol length, generally, L can meet the requirement by taking 64, and arg represents angle taking operation;

and when the number of the output data reaches LN, estimating the initial phase of the clock.

4. The timing synchronization method of initial phase injection according to claim 1, further comprising the steps of, in step S2:

s21: the input data is calculated to obtain an interpolation sequence of a loop through an interpolation filter under the control of a second numerical control oscillator;

s22: and performing timing error estimation operation on the interpolation sequence according to the Gardner algorithm to obtain a timing error signal.

5. The timing synchronization method of initial phase injection according to claim 1, further comprising the steps of, in step S3:

s31: inputting the timing error into a loop filter to filter out high frequency components, and simultaneously inputting the timing error into a gain control filter;

s32: the selection controller judges the loop state by calculating the error average value of the timing error, thereby adjusting the loop coefficient;

s33: the result of multiplying the sum of the loop filter and the gain control filter output by the loop coefficient is used as the control word for the second numerically controlled oscillator.

6. The timing synchronization method of initial phase injection according to claim 1, further comprising the steps of, in step S4:

s41: the second numerically controlled oscillator determines interpolation base points and fractional intervals required by the interpolation filter in an overflow counting mode;

s42: the interpolation filter completes a new round of interpolation calculation by utilizing the interpolation base points and the fractional intervals, thereby forming a closed-loop feedback loop, and continuously adjusting the positions of the interpolation base points through cyclic iteration to obtain optimal sampling data.

Images