CN114785475B - Enhancement method for Gardner timing error detection - Google Patents
Enhancement method for Gardner timing error detection Download PDFInfo
- Publication number
- CN114785475B CN114785475B CN202210374052.XA CN202210374052A CN114785475B CN 114785475 B CN114785475 B CN 114785475B CN 202210374052 A CN202210374052 A CN 202210374052A CN 114785475 B CN114785475 B CN 114785475B
- Authority
- CN
- China
- Prior art keywords
- timing error
- adjacent
- symbols
- gardner
- symbol
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L7/00—Arrangements for synchronising receiver with transmitter
- H04L7/0054—Detection of the synchronisation error by features other than the received signal transition
Abstract
The application discloses an enhancement method for Gardner timing error detection, which directly removes or reduces self-noise interference from different sources in the traditional Gardner method. For BPSK/QPSK equivalent baseband signals filtered by adopting low roll-off coefficients, when adjacent symbols have polarity jump, firstly removing the influence of the front and rear 2 symbols which are adjacent to the adjacent symbols on an intermediate sampling point, and then using the intermediate sampling point for calculating timing errors; when the adjacent symbol has no polarity jump, the timing error is set to zero directly. By directly removing or reducing the self-noise interference of the band-limited signal, the application not only can expand the applicable band-limited range of the traditional Gardner timing error detection method, but also is beneficial to improving the acquisition time of the timing loop and the anti-noise performance thereof.
Description
Technical Field
The application belongs to the technical field of synchronization of digital communication, and particularly relates to an enhancement method for Gardner timing error detection.
Background
Timing synchronization is sometimes also referred to as symbol timing synchronization or symbol synchronization. As the name implies, symbol timing synchronization is a process of aligning symbol clocks at a receiving end with corresponding clocks at a transmitting end in a synchronous manner, and the purpose is to overcome the sliding code phenomenon which may occur in long-term operation of the receiving end due to clock crystal oscillator drift and the like; and secondly, judging the received symbol at the moment of maximum signal-to-noise ratio so as to obtain the optimal receiving performance.
The conventional Gardner timing error detection method has a simple structure, requires only 2 sampling points (i.e., decision sample points and intermediate sample points) for each symbol, is insensitive to carrier phase, and thus is widely used in the timing loop of a digital receiver, but is susceptible to self-noise. Because raised cosine filters are commonly used to constrain signal spectrum and control inter-symbol interference in band-limited signal systems today. However, as the roll-off coefficient of the raised cosine filter is reduced, the jitter of the filtered signal between the optimal sampling points becomes large, which results in a larger variance of the output result of the conventional Gardner timing error detection method. In this case, if the timing loop bandwidth is reduced in order to reduce the timing phase jitter, the lock time of the timing loop tends to be increased. Therefore, the raised cosine roll-off coefficient suitable for the traditional Gardner timing error detection method is between 0.4 and 1, which is not beneficial to improving the frequency band utilization rate.
Disclosure of Invention
The application aims to: in order to overcome the defects in the prior art, the application provides an enhancement method for Gardner timing error detection, which effectively reduces self-noise interference caused by the reduction of raised cosine roll-off coefficient.
The technical scheme is as follows: the application discloses a Gardner timing error detection enhancement method, for BPSK (Binary Phase Shift Keying ) and/or QPSK (Quadrature Phase Shift Keying, quadrature phase shift keying) equivalent baseband signals subjected to raised cosine filtering by adopting a low roll-off coefficient, when polarity jump exists in adjacent symbols, the influence of N symbols immediately adjacent to the adjacent symbols on a middle sampling point is removed, N is an even number larger than or equal to 2, a corrected middle sampling point is obtained, and timing error is calculated according to the corrected middle sampling point; when the adjacent symbol has no polarity jump, the timing error is set to zero directly.
Further, when the adjacent symbol has polarity jump, the influence of the front and rear 2 symbols adjacent to the adjacent symbol on the middle sampling point is removed, namely, the influence of the front and rear 2 symbols which are positioned at two sides of the middle sampling point and are separated by 3/2 symbol periods on the middle sampling point is removed, and then the middle sampling point is used for calculating the timing error.
Further, the decision samples of adjacent symbols are noted as y (t k-1 ) And y (t) k ) The first and second 2 symbols immediately adjacent to the adjacent symbol are y (t k-2 ) And y (t) k+1 ) The method comprises the steps of carrying out a first treatment on the surface of the The effect of the next 2 symbols immediately adjacent to the adjacent symbol on the intermediate samples is β (y (t k-2 )+y(t k+1 ) Where β represents the weight coefficient of the smearing.
Further, timing error when there is a polarity jump in adjacent symbols of the in-phase branch Wherein (1)>Representing the real part of {.cndot }, y (t k-1/2 ) Intermediate samples representing adjacent symbols; otherwise, when adjacent symbols of the in-phase branch are notForced timing error e when polarity jump exists I (t k )＝0。
Further, timing error when there is a polarity jump in adjacent symbols of the orthogonal branch Wherein (1)>An imaginary part representing { · }; otherwise, when the adjacent symbols of the orthogonal branch do not have polarity jump, the timing error e is forced Q (t k )＝0。
Further, in the case where the roll-off coefficient is known, a weight coefficient β=g (3T/2)/g (0) of the smear disturbance is calculated, where g (kT) is the impulse response of the raised cosine roll-off filter and T is the symbol period.
Further, the low cosine roll-off coefficient refers to a cosine roll-off coefficient of 0.4 or less.
Further, the timing error in the symbol period T is e (T k )＝e I (t k )+e Q (t k )。
Wherein y is I [·]And y Q [·]Respectively represent BPSK/QPSK equivalent baseband signals y [. Cndot.]Is used for the real and imaginary parts of (a),representing a decision value or a sign function.
Further, the enhancement method is applied to a timing error detector of a digital phase-locked loop structure, and calibration is performed according to the timing error.
Further, the timing error detector employs a 7-stage shift memory to shift-store the BPSK/QPSK equivalent baseband signal samples input at a 2-fold symbol rate.
The beneficial effects are that:
the application can not only enlarge the applicable band limit range of the traditional Gardner timing error detection method, but also help to improve the capturing time of the timing loop and the noise resistance thereof by directly removing or reducing the self-noise interference of the band limit signal.
Drawings
The foregoing and/or other advantages of the application will become more apparent from the following detailed description of the application when taken in conjunction with the accompanying drawings and detailed description.
Fig. 1 is a block diagram of a timing synchronization loop.
Fig. 2 is a BPSK signal eye diagram with a roll-off coefficient of 1.
Fig. 3 is a BPSK signal eye diagram with a roll-off coefficient of 0.5.
Fig. 4 is a BPSK signal eye diagram with a roll-off coefficient of 0.25.
Fig. 5 is a raised cosine pulse waveform of different roll-off coefficients.
Fig. 6 is a block flow diagram of enhanced Gardner timing error detection.
Detailed Description
Embodiments of the present application will be described below with reference to the accompanying drawings.
The enhancing method for Gardner timing error detection provided by the embodiment of the application is applied to a band-limited signal system using a raised cosine filter, wherein the band-limited signal comprises but is not limited to a BPSK/QPSK equivalent baseband signal. For example, in a typical digital phase-locked loop architecture such as that of fig. 1, the typical digital phase-locked loop architecture of fig. 1 includes an interpolator, a timing error detector, a loop filter, and a timing controller. The enhancement method for Gardner timing error detection provided by the embodiment of the application is applied to the timing error detector, and is also called an enhancement Gardner timing error detector; the application is not modified from other components except for the timing error detector, and is completely consistent with the traditional Gardner method.
Since the conventional Gardner timing error detection algorithm requires 2 sample points per symbol, the output rate of the interpolator should be 2 times the symbol rate. Also for ease of implementation, the input rate of the interpolator is typically set to 2 times the output rate, i.e. 4 times the symbol rate. In fig. 1, the input of the matched filter is r (t), and the output of the matched filter is y (t); sampling the matched filter result y (t) at a fixed rate achieves the input rate configuration of the interpolator.
In the timing error detector, every 2 samples { y (t) k-1/2 )，y(t k ) Calculating 1 st time timing error e (t) k ). Seemingly, the timing error detector needs to determine the sequence of input of the sample points and the intermediate sample points clearly so as to calculate the timing error correctly. However, due to the automatic adjustment function of the feedback loop, even if the actual sample input sequence at the start of the loop does not coincide with the input sequence assumed by the timing error detector, the actual sample input sequence is automatically adjusted to the input sequence assumed by the timing error detector when the loop is finally locked.
Loop filters typically employ an active proportional-integral filter to correct the timing error e (t k ) The step value of the timing controller is adjusted after filtering. The step basic value of the timing controller is fixed at 0.5, and the input-output basic rate ratio of the interpolator is controlled. The loop-filtered timing error is used as a step offset value for the timing controller, and the combined step base value updates its internal modulo-1 down counter at 4 times the symbol rate. Each time the counter underflow occurs, a fractional time interval is calculated from the content before the counter underflow. The interpolator calculates a corrected interpolation value based on the sampling time at which the counter has underflow and the fractional time interval.
The conventional Gardner method calculates the timing error e (t) k ) The method of (2) is as follows:
e(t k )＝{y I (t k-1 )-y I (t k )}·y I (t k-1/2 )+{y Q (t k-1 )-y Q (t k )}·y Q (t k-1/2 ) (1)
wherein y is I (t k ) And y Q (t k ) Sample points y (t) of BPSK/QPSK equivalent baseband signals respectively k ) Is the real and imaginary parts of (c), y (t k-1 ) And y (t) k ) Representing two adjacent decision samples, y (t k-1/2 ) Are their intermediate spots. The conventional Gardner method uses decision samples to generate direction information and multiplies the direction information with intermediate samples to obtain a measure of timing error. It can be seen that the nature of the conventional Gardner method, i.e., zero crossing detection, significantly reduces the accuracy of the timing error metric if the intermediate samples are disturbed by noise.
It is well known that the conventional Gardner method is susceptible to self-noise even if external noise is not considered. This is because, in band-limited signal systems, raised cosine filters are generally used to constrain the signal spectrum and control inter-symbol interference, with lower raised cosine filters having higher band-limited coefficients. However, as the roll-off coefficient decreases, there is no intersymbol interference at the optimal sampling points, but the jitter of the signal between the optimal sampling points becomes large. As shown by the BPSK signal eye diagrams for the different roll-off coefficients in fig. 2, 3 and 4. It can be seen that the smaller the roll-off coefficient, the greater the self-noise at the intermediate sample point. When the roll-off coefficient is reduced from 1 to 0.25, the middle sample point also gradually spreads upwards and downwards from the 1 zero crossing point which is converged initially, and the maximum amplitude can even reach 0.64 of the decision sample point, so that the roll-off coefficient suitable for the traditional Gardner method is between 0.4 and 1.
The raised cosine pulse waveforms of different roll-off coefficients in fig. 5 are observed, the smaller the roll-off coefficient is, the slower the trailing decay of the pulse waveform is. It is envisaged that when adjacent symbols y (t k-1 ) And y (t) k ) When there is no polarity jump, there is no zero crossing in the middle, so timing error information cannot be provided. Only if adjacent symbols y (t k-1 ) And y (t) k ) The intermediate zero crossing point can provide timing error information when a polarity jump occurs. At this time, the middle sample point y (t k-1/2 ) Is not from two side symbols y (t) k-1 ) And y (t) k ) But from other symbols than the two-sided symbol. That is, the pulseThe main lobe of the punching waveform is the main source of self-noise at the middle sampling point, but no zero crossing point exists at the moment; or zero crossings occur, but hardly self-noise is generated because of the mutual cancellation. When the main lobes cancel each other at the middle sample point, the trailing oscillations will become the main source of self-noise.
The embodiment of the application provides an enhancement method for Gardner timing error detection, which is characterized in that for BPSK/QPSK equivalent baseband signals filtered by low cosine roll-off coefficients, when polarity jump exists in adjacent symbols, the influence of N symbols immediately adjacent to the adjacent symbols on middle sampling points is removed, N is an even number larger than or equal to 2, corrected middle sampling points are obtained, and timing errors are calculated according to the corrected middle sampling points; when the adjacent symbol has no polarity jump, the timing error is set to zero directly.
In this embodiment, if only the influence of trailing oscillation in the 1 st symbol period from both sides of the main lobe is considered, firstly removing the influence of the front and rear 2 symbols immediately adjacent to the adjacent symbol on the middle sampling point, that is, firstly removing the influence of the front and rear 2 symbols spaced 3/2 symbol periods on both sides of the middle sampling point on the middle sampling point, correcting the middle sampling point according to the influence on the middle sampling point, and then calculating the timing error according to the corrected middle sampling point, the specific process includes the following steps:
step 1, in the enhanced Gardner timing error detector provided in this embodiment, 7-level shift storage is performed on BPSK/QPSK equivalent baseband signal samples input at 2 times of symbol rate, and are respectively labeled as y (t k-2 )、y(t k-3/2 )、y(t k-1 )、y(t k-1/2 )、y(t k )、y(t k+1/2 ) And y (t) k+1 ) A total of 7 signal samples are consecutive.
Step 2, every two signal samples (i.e. 1 symbol period T) are input, and the sample is taken out in step l and marked as y (T k-2 )、y(t k-1 )、y(t k-1/2 )、y(t k ) And y (t) k+1 ) Is a single sample of the signal. Wherein y (t) k-2 )、y(t k-1 )、y(t k ) And y (t) k+1 ) As decision samples, y (t k-1/2 ) As an intermediate sample.
Step 3, removing the 2 outermost decision samples y (t k-2 ) And y (t) k+1 ) For the middle sample y (t) k-1/2 ) And then multiplies the influence of the intermediate 2 adjacent decision samples y (t k-1 ) And y (t) k ) The difference is calculated to obtain the current timing error e (t k )。
e(t k )＝e I (t k )+e Q (t k )
Wherein y is I [·]And y Q [·]Respectively represent BPSK/QPSK equivalent baseband signals y [. Cndot.]Is used for the real and imaginary parts of (a),represents a decision value or a sign function, β represents a weight coefficient of trailing interference, β (y (t k-2 )+y(t k+1 ) I.e. the effect of the front and rear 2 symbols spaced 3/2 symbol periods on the middle sample on the two sides of the middle sample. In the case where the roll-off coefficient is known, the weight coefficient β=g (3T/2)/g (0) of the smearing could be calculated in advance, where g (kT) is the impulse response of the raised cosine roll-off filter and T is the symbol period. The low cosine roll-off coefficient refers to a cosine roll-off coefficient less than or equal to 0.4. For example, when the roll-off coefficient takes a value of 0.25, the weight coefficient β= -0.1856 of the trailing interference at this time is easy to calculate.
As with the conventional Gardner method, the enhanced Gardner method also requires 2 sample points per symbol, i.e., a 2 times symbol rate input. But different, the enhancement method requires 4 decisions at the same time per 1 timing error calculatedSpots and 1 intermediate spot. That is, at least 7 stages of shift memories should be used to shift-store the BPSK/QPSK equivalent baseband signal samples input at 2 times the symbol rate. Assume that, in chronological order, they are denoted as y (t k-2 )、y(t k-3/2 )、y(t k-1 )、y(t k-1/2 )、y(t k )、y(t k+1/2 ) And y (t) k+1 ) A total of 7 complex signal samples.
Every 2 samples are input, i.e. take out the label y (t k-2 )、y(t k-1 )、y(t k-1/2 )、y(t k ) And y (t) k+1 ) Is a single sample of 5 samples. Wherein y (t) k-2 )、y(t k-1 )、y(t k ) And y (t) k+1 ) Are all regarded as decision samples, y (t k-1/2 ) As an intermediate sample. Then the outermost 2 decision samples y (t) are removed according to equation (2) k-2 ) And y (t) k+1 ) For the middle sample y (t) k-1/2 ) And calculates the current timing error e (t k )。
The procedure flow for enhanced timing error detection is shown in fig. 6. Wherein the input flag is used to control every 2 samples input, enabling 1 timing error calculation. Although the input flag value is related to the sample input order assumed by the timing error detector, whether the loop eventually converges is independent of the initial value of the input flag due to the automatic adjustment function of the feedback loop. Of course, in applications requiring rapid convergence, such as burst communication, other auxiliary methods may be discussed to distinguish the sequence of inputting the decision sample point and the intermediate sample point in advance, and then particularly set the input flag value to accelerate the convergence speed of the timing synchronization loop.
The present application provides a method for enhancing Gardner timing error detection, and the method and means for implementing the technical scheme are numerous, and the above description is only a specific embodiment of the present application, and it should be noted that, for those skilled in the art, several improvements and modifications can be made, and these improvements and modifications should also be considered as the protection scope of the present application. The components not explicitly described in this embodiment can be implemented by using the prior art.
Claims (5)
1. A method for enhancing Gardner timing error detection, comprising: for BPSK/QPSK equivalent baseband signals filtered by adopting low cosine roll-off coefficients, when polarity jump exists in adjacent symbols, firstly removing the influence of the front and rear 2 symbols which are adjacent to the adjacent symbols on the middle sampling point, namely firstly removing the influence of the front and rear 2 symbols which are positioned at two sides of the middle sampling point and have 3/2 symbol periods on the middle sampling point, obtaining a corrected middle sampling point, and then calculating a timing error according to the corrected middle sampling point; when the adjacent symbol does not have polarity jump, the timing error is directly set to zero;
the decision samples of adjacent symbols are noted as y (t k-1 ) And y (t) k ) The first and second 2 symbols immediately adjacent to the adjacent symbol are y (t k-2 ) And y (t) k+1 ) The method comprises the steps of carrying out a first treatment on the surface of the The effect of the next 2 symbols immediately adjacent to the adjacent symbol on the intermediate samples is β (y (t k-2 )+y(t k+1 ) Where β represents the weight coefficient of the trailing interference;
timing error when there is a polarity jump in adjacent symbols of the in-phase branch Wherein (1)>Representing the real part of {.cndot }, y (t k-1/2 ) Intermediate samples representing adjacent symbols; otherwise, when the adjacent symbols of the in-phase branch do not have polarity jump, the timing error e is forced I (t k )＝0；
Timing error when there is a polarity jump in adjacent symbols of orthogonal branches Wherein (1)>An imaginary part representing { }; otherwise, when the adjacent symbols of the orthogonal branch do not have polarity jump, the timing error e is forced Q (t k )＝0；
The timing error in the symbol period T is e (T k )＝e I (t k )+e Q (t k )。
2. An enhanced method of Gardner timing error detection according to claim 1 wherein: in case the roll-off coefficient is known, a weight coefficient β=g (3T/2)/g (0) of the trailing interference is calculated, where g (kT) is the impulse response of the raised cosine roll-off filter and T is the symbol period.
3. An enhanced method of Gardner timing error detection according to claim 2 wherein: the low cosine roll-off coefficient refers to a cosine roll-off coefficient less than or equal to 0.4.
4. A method for enhancing Gardner timing error detection according to claim 3 wherein: and the timing error detector is applied to the digital phase-locked loop structure and is calibrated according to the timing error.
5. The method for enhancing Gardner timing error detection according to claim 4, wherein: the timing error detector employs a 7-stage shift memory to shift-store the BPSK/QPSK equivalent baseband signal samples input at a 2-fold symbol rate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210374052.XA CN114785475B (en) | 2022-04-11 | 2022-04-11 | Enhancement method for Gardner timing error detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210374052.XA CN114785475B (en) | 2022-04-11 | 2022-04-11 | Enhancement method for Gardner timing error detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114785475A CN114785475A (en) | 2022-07-22 |
| CN114785475B true CN114785475B (en) | 2023-09-22 |
Family
ID=82428496
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210374052.XA Active CN114785475B (en) | 2022-04-11 | 2022-04-11 | Enhancement method for Gardner timing error detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114785475B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH1028149A (en) * | 1996-07-10 | 1998-01-27 | Matsushita Electric Ind Co Ltd | Clock recovery circuit |
| CN103457680A (en) * | 2013-08-20 | 2013-12-18 | 重庆邮电大学 | Satellite communication timing synchronization error detection method based on full-digital receiving |
| CN104065603A (en) * | 2014-06-14 | 2014-09-24 | 温州大学 | Quadrature phase shift keying signal automatic identification method in raised cosine sending moulding and multipath environment |
-
2022
- 2022-04-11 CN CN202210374052.XA patent/CN114785475B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH1028149A (en) * | 1996-07-10 | 1998-01-27 | Matsushita Electric Ind Co Ltd | Clock recovery circuit |
| CN103457680A (en) * | 2013-08-20 | 2013-12-18 | 重庆邮电大学 | Satellite communication timing synchronization error detection method based on full-digital receiving |
| CN104065603A (en) * | 2014-06-14 | 2014-09-24 | 温州大学 | Quadrature phase shift keying signal automatic identification method in raised cosine sending moulding and multipath environment |
Non-Patent Citations (2)
| Title |
|---|
| 一种改进的Gardner定时同步算法;刘伟;姚远程;秦明伟;;计算机工程(11);全文 * |
| 用于减小定时抖动的数字预滤波器设计;孔明东, 邱昆;电子科技大学学报(06);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114785475A (en) | 2022-07-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2170344C (en) | Signal processing system | |
| CN101299657B (en) | Symbol timing synchronizing apparatus for complete digital receiver | |
| US5870438A (en) | Fast resynchronization system for high-speed data transmission | |
| US6985549B1 (en) | Blind cost criterion timing recovery | |
| EP2903199B1 (en) | Circuit for symbol timing synchronization | |
| US5838744A (en) | High speed modem and method having jitter-free timing recovery | |
| CN106603217B (en) | Sampling frequency offset suppression method for Bluetooth signal of wireless comprehensive tester | |
| US6690753B2 (en) | Receiver having decisional feedback equalizer with remodulation and related methods | |
| CN111130595B (en) | Low-earth-orbit satellite feed link timing synchronization method | |
| US4831637A (en) | Apparatus and technique for timing jitter cancellation in a data receiver | |
| JP4389934B2 (en) | Clock recovery circuit | |
| US6154510A (en) | Symbol timing recovery based on adjusted, phase-selected magnitude values | |
| US7660377B2 (en) | Device for estimating a timing correction loop error for a digital demodulator | |
| CN114785475B (en) | Enhancement method for Gardner timing error detection | |
| US7233632B1 (en) | Symbol timing correction for a phase modulated signal with mutually interfering symbols | |
| EP3456009B1 (en) | Timing recovery with adaptive channel response estimation | |
| CN114465691A (en) | Low-complexity constant envelope phase modulation signal sampling deviation estimation and compensation method and system | |
| US7463701B2 (en) | Fast symbol timing recovery for transmitted digital data | |
| JP4335125B2 (en) | Timing synchronization circuit | |
| US7263139B1 (en) | Phase correction for a phase modulated signal with mutually interfering symbols | |
| Leng et al. | A modified gardner detector for multilevel PAM/QAM system | |
| CN111106925B (en) | Timing synchronization method for reducing steady-state error | |
| CN113542166B (en) | Timing recovery method and device with low jitter and rapid convergence | |
| JP2001211221A (en) | Timing phase acquisition method and device for remote communication system | |
| JP2000183992A (en) | Clock reproducing method and circuit |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |