A symbol synchronization method and device based on signal-to-noise ratio optimization

Abstract

This invention discloses a symbol synchronization method and apparatus based on signal-to-noise ratio (SNR) optimization, relating to the field of digital trunking communication technology. It aims to address the insufficient accuracy of correlation peak-based synchronization methods. The method includes the following steps: S1, receiving a radio frequency (RF) signal, processing it, and then sampling it to obtain a digital sampling sequence; S2, performing a cross-correlation operation between the digital sampling sequence and a locally stored synchronization word, and using the maximum value of the found correlation peak as the initial synchronization position; S3, obtaining the optimal symbol synchronization position near the initial synchronization position through soft decision and SNR calculation; S4, demodulating the data payload at the optimal symbol synchronization position to complete the current frame processing. This invention can achieve both fast acquisition using a synchronization word and high-precision symbol timing.

Description

Technical Field

[0001] This invention relates to the field of digital trunking communication technology, and specifically to a symbol synchronization method and apparatus based on signal-to-noise ratio optimization. Background Technology

[0002] Digital trunking communication systems, especially the PDT (Dedicated Digital Trunking) standard, play a crucial role in modern emergency command and public safety. Their communication quality and reliability are highly dependent on the synchronization performance of the receiver. Symbol synchronization (or bit synchronization) is a key step, aiming to accurately determine the optimal sampling time for each symbol at the receiver to maximize the signal-to-noise ratio and reduce the bit error rate.

[0003] Currently, existing symbol synchronization technologies mainly suffer from the following limitations. Synchronization based on correlation peaks is the most commonly used method, such as the Chinese patent with publication number CN102223345A. The receiver calculates the cross-correlation function between the received signal and a locally known synchronization word, and finds the location of the maximum value of the correlation peak to determine the approximate starting boundary of the symbol. However, this method has inherent drawbacks: 1. Limited accuracy: Due to channel distortion, noise, and multipath effects, the correlation peak often becomes flat or distorted, resulting in the found maximum value not being the theoretically optimal sampling point. The optimal sampling point should be located at the position where the eye diagram is widest within the symbol period, a point that the correlation peak method cannot directly find; 2. Dependence on synchronization word: This method is only effective in frames containing a synchronization word. For subsequent voice data frames that do not contain a synchronization word, the system usually relies on circuits such as phase-locked loops for blind synchronization. Once synchronization is lost due to clock drift or channel abrupt changes, it needs to wait for the next synchronization word frame to re-acquire, resulting in data loss during this period.

[0004] Other blind synchronization methods include those using signal envelopes and zero-crossing detection. These methods typically do not rely on specific data sequences, but they are less robust to noise and their performance degrades sharply in low signal-to-noise ratio environments. Their synchronization accuracy and stability are generally inferior to data-assisted synchronization methods. Summary of the Invention

[0005] This invention addresses the problem of insufficient accuracy in synchronization methods based on correlation peaks by proposing a symbol synchronization method and apparatus based on signal-to-noise ratio optimization. This method can achieve both rapid acquisition using synchronization words and high-precision symbol timing.

[0006] To achieve the above objectives, the present invention adopts the following technical solution, wherein the first technical solution is a symbol synchronization method based on signal-to-noise ratio optimization, comprising the following steps: S1 receives radio frequency signals, processes them, and then samples them to obtain a digital sampling sequence. S2, perform cross-correlation operation between the digital sampling sequence and the locally stored synchronization word, and take the maximum value of the found correlation peak as the initial synchronization position; S3, through soft decision and SNR calculation, obtains the optimal symbol synchronization position near the initial synchronization position; S4, demodulate the data payload at the optimal symbol synchronization position to complete the processing of the current frame.

[0007] The technical solution of the present invention includes the following process: first, signal reception and sampling are performed, followed by coarse synchronization of correlation peaks, and after coarse synchronization is completed, fine synchronization of signal-to-noise ratio is performed. The fine synchronization of signal-to-noise ratio is the core step of the present invention. After fine synchronization of signal-to-noise ratio is completed, data demodulation is performed.

[0008] The present invention is further configured such that step S3 includes the following steps: S31, Select several sampling points near the initial synchronization position as candidate points; S32, perform soft decision and SNR calculation based on the above candidate points; S33. Select the position corresponding to the candidate point with the largest SNR value as the final optimal symbol synchronization position.

[0009] In this technical solution, for fine synchronization of signal-to-noise ratio, a candidate point set is first determined, then soft decision and SNR calculation are performed based on the candidate points, and finally the optimal synchronization position is selected by comparison based on the calculation results.

[0010] The present invention is further configured such that step S1 includes: after the radio frequency signal is down-converted and filtered, a baseband signal is obtained, and the baseband signal is sampled by an analog-to-digital converter at a frequency several times that of the symbol rate to obtain a digital sampling sequence.

[0011] The symbol rate in this technical solution can be 20 times the symbol rate.

[0012] The present invention is further configured to include step S5, wherein step S5 includes the synchronization maintenance and fine-tuning of the unsynchronized word frame, wherein the synchronization maintenance process of the unsynchronized word frame includes: using the clock phase corresponding to the optimal synchronization position P_optimal determined in the previous frame for maintenance, as the initial synchronization position P_maintain of the current frame.

[0013] The present invention is further configured such that the fine-tuning process of the unsynchronized word frame includes: selecting candidate points near the initial synchronization position P_maintain of the current frame, calculating the SNR value of each point, and selecting the point with the highest SNR as the new synchronization position after fine-tuning.

[0014] In this technical solution, the synchronization state established in the previous frame is used for maintenance, and the signal-to-noise ratio comparison mechanism is also applied for fine-tuning, so as to achieve long-term stable synchronization even without a synchronization word.

[0015] The present invention is further configured such that: the soft decision includes: using a candidate position as the symbol starting point, demodulating the entire current data frame to obtain a soft information sequence for each symbol.

[0016] In this technical solution, the soft information can be the amplitude, phase, or likelihood ratio information output by the demodulator.

[0017] The present invention is further configured such that the SNR calculation includes: S321, Generate an ideal reference signal sequence, perform hard decision on each value xs_i in the soft information sequence, and map it to the nearest standard modulation symbol value; S322, calculate the average value of the mapping results respectively, calculate the scaling factor k based on the average value, and obtain the sequence S_s by multiplying the scaling factor k and the standard modulation symbol value; S323 calculates the total energy and noise energy, obtains the signal energy based on the total energy and noise energy, and finally estimates the signal-to-noise ratio.

[0018] The present invention is further configured such that: the soft information sequence X_s is represented as [xs_1, xs_2, ..., xs_N], and the ideal reference signal sequence S_s is represented as [ss_1, ss_2, ..., ss_N], where N is the frame length; and the closest standard modulation symbol value is +3, +1, -1, -3.

[0019] In this technical solution, an ideal reference signal sequence S_s is generated in order to estimate the noise.

[0020] The present invention is further configured such that step S322 includes: Take out the xs_i with hard decision results of +3 and -3 respectively, and calculate their average values ​​mean_3 and mean_-3; Take out the xs_i with hard decision results of +1 and -1 respectively, and calculate their average mean_1 and mean_-1; The scaling factor k is the sum of mean_3 and mean_1 minus the sum of mean_-1 and mean_-3, then divided by 8.

[0021] In this technical solution, the scaling factor k reflects the combined effect of channel gain and distortion on the signal amplitude.

[0022] The second approach is a symbol synchronization device based on signal-to-noise ratio (SNR) optimization, applicable to the aforementioned symbol synchronization method based on SNR optimization. It includes a correlation peak synchronization module and an SNR calculation and selection module connected to the correlation peak synchronization module. The SNR calculation and selection module is also connected to a demodulation module, and the demodulation module is also connected to a synchronization maintenance and fine-tuning module.

[0023] In this technical solution, the correlation peak synchronization module, the signal-to-noise ratio calculation and selection module, the demodulation module, and the synchronization maintenance and fine-tuning module are connected in sequence to form a symbol synchronization device based on signal-to-noise ratio optimization. This device can effectively overcome the synchronization error caused by noise and multipath effects, and improve the demodulation performance and decoding success rate of the PDT system.

[0024] The present invention can bring the following beneficial effects: This invention relates to a symbol synchronization method based on signal-to-noise ratio (SNR) optimization. Unlike traditional methods that rely solely on the shape of correlation peaks, this method innovatively uses SNR, a direct indicator of communication link quality, as the final criterion for determining synchronization position. By selecting the sampling point with the highest SNR, the optimal sampling time with the strongest anti-interference capability and the largest eye diagram opening can be accurately located, fundamentally improving synchronization accuracy. Attached Figure Description

[0025] Figure 1 This is a flowchart of a symbol synchronization method based on signal-to-noise ratio optimization according to the present invention.

[0026] Figure 2 This is a schematic diagram of a symbol synchronization device based on signal-to-noise ratio optimization according to the present invention. Detailed Implementation

[0027] Example 1 To address the insufficient accuracy of existing correlation peak-based synchronization methods, this embodiment proposes a symbol synchronization method based on signal-to-noise ratio optimization, referencing... Figure 1 It mainly includes the following steps.

[0028] Step S1: First, receive the corresponding radio frequency signal, process the radio frequency signal accordingly, and then sample to obtain a digital sampling sequence.

[0029] More specifically, in step S1 above, after receiving the radio frequency signal, it undergoes down-conversion and filtering to obtain the corresponding baseband signal, which is then sampled by the ADC analog-to-digital converter at a frequency several times the symbol rate, ultimately yielding a digital sampling sequence.

[0030] In this embodiment, sampling is performed at 20 times the symbol rate.

[0031] Down-conversion shifts high-frequency radio frequency signals to more easily processed frequencies. Filtering removes high-frequency components and noise generated during down-conversion; specifically, bandpass filters can be used for this purpose.

[0032] Step S2: Then, perform cross-correlation operation between the digital sampling sequence and the locally stored synchronization word, and take the maximum value of the found correlation peak as the initial synchronization position.

[0033] Step S2 is the coarse synchronization process for the correlation peak. It performs cross-correlation calculations on the sampled data obtained in step S1 and the locally stored synchronization word. By scanning the entire calculation result, the maximum value point of the correlation peak can be found. The sampling position corresponding to this maximum value point is marked as the initial synchronization position P0, thereby completing frame synchronization and coarse symbol synchronization.

[0034] After completing step S2, proceed to step S3, and obtain the optimal symbol synchronization position near the initial synchronization position through soft decision and SNR calculation.

[0035] In this embodiment, the coarse synchronization position P0 obtained from the coarse synchronization process of the correlation peak in step S2 is not the optimal sampling point. It is necessary to find the optimal sampling point with the highest signal-to-noise ratio near the coarse synchronization position P0.

[0036] More specifically, it mainly includes the following sub-steps, including three processes: determining candidate sampling points, soft decision and SNR calculation, and comparing and selecting the optimal synchronization position.

[0037] Step S31: Select several sampling points near the initial synchronization position as candidate points.

[0038] More specifically, for step S31 above, a candidate sampling point set is determined by selecting the sampling points adjacent to the coarse synchronization position P0 as candidate points.

[0039] In this embodiment, three candidate points are selected: {P0-1, P0, P0+1}.

[0040] Step S32: Perform soft decision and SNR calculation based on the above candidate points.

[0041] For each candidate position, the soft decision and SNR calculation process are performed sequentially.

[0042] In this embodiment, the candidate position P_candidate is used as an example for explanation.

[0043] The soft decision process includes the following steps: starting from a candidate position, demodulate the entire current data frame to obtain the soft information sequence for each symbol.

[0044] In this embodiment, P_candidate is used as the symbol starting point, and the entire current data frame (containing known synchronization words and unknown data payloads) is demodulated to obtain the soft signal sequence X_s for each symbol, which is represented as [xs_1, xs_2, ..., xs_N], where N is the frame length; where the soft information can be the amplitude, phase or likelihood ratio information output by the demodulator.

[0045] SNR calculation mainly includes the following process.

[0046] Step S321: Generate an ideal reference signal sequence, perform a hard decision on each value xs_i in the soft information sequence, and map it to the closest standard modulation symbol value.

[0047] Specifically, in order to estimate the noise, an ideal reference signal sequence S_s needs to be generated, which is represented as [ss_1, ss_2, ..., ss_N], where N is the frame length.

[0048] Then, a hard decision is made on each value xs_i in the soft signal sequence X_s, mapping it to the nearest standard modulation symbol value.

[0049] In this embodiment, for 4FSK, the standard values ​​are +3, +1, -1, -3.

[0050] Step S322: Calculate the average value of the mapping results respectively, calculate the scaling factor k based on the average value, and obtain the sequence S_s by multiplying the scaling factor k and the standard modulation symbol value.

[0051] More specifically, the xs_i with hard decision results of +3 and -3 are extracted respectively, and their average values ​​mean_3 and mean_-3 are calculated. The xs_i with hard decision results of +1 and -1 are extracted respectively, and their average values ​​mean_1 and mean_-1 are calculated.

[0052] The scaling factor k is equal to the sum of mean_3 and mean_1 minus the sum of mean_-1 and mean_-3, and then divided by 8; this scaling factor reflects the combined effect of channel gain and distortion on signal amplitude.

[0053] The reference sequence is then generated by multiplying the standard symbol value (+3, +1, -1, -3) to be decided by the scaling factor k to obtain each value ss_i in the reference information sequence S_s, that is, ss_i is equal to the hard decision result multiplied by the scaling factor k; this reference signal sequence S_s can be regarded as the soft information that should be obtained "ideally" after passing through the same channel distortion.

[0054] Step S323: Calculate the total energy and noise energy, obtain the signal energy based on the total energy and noise energy, and finally estimate the signal-to-noise ratio.

[0055] The calculation process of signal-to-noise ratio (SNR) mainly includes the total energy calculation process, noise energy calculation process, signal energy calculation process, and SNR estimation process.

[0056] For the calculation of total energy: the total energy E_total is equal to the sum of the squares of the absolute values ​​of the signal samples xs_i; where i ranges from 1 to N.

[0057] For noise energy calculation: Noise energy E_noise is equal to the sum of the squares of the absolute values ​​of the differences between xs_i and ss_i, where i ranges from 1 to N. This calculation quantifies the difference between the received signal and the ideal reference signal, i.e., noise and distortion.

[0058] For signal energy calculation: signal energy E_signal is the difference between the total energy E_total and the noise energy E_noise.

[0059] Signal-to-noise ratio estimation: The signal-to-noise ratio SNR_est_dB is equal to 10 multiplied by the common logarithm of the ratio of signal energy E_signal to noise energy E_noise.

[0060] Step S33: Select the position corresponding to the candidate point with the largest SNR value as the final optimal symbol synchronization position.

[0061] In more detail, the optimal synchronization positions are compared and selected. The SNR_est_dB values ​​for the three candidate positions P0-1, P0, and P0+1 are calculated respectively. The position corresponding to the candidate point with the largest SNR value is selected as the final optimal symbol synchronization position P_optimal.

[0062] Step S4: Demodulate the data payload at the optimal symbol synchronization position to complete the processing of the current frame.

[0063] At the optimal symbol synchronization position P_optimal, the subsequent data payload is sampled and decided to complete the demodulation of the data.

[0064] Example 2 This embodiment proposes a symbol synchronization method based on signal-to-noise ratio optimization, referring to... Figure 1 It mainly includes the following steps.

[0065] Step S1: First, receive the corresponding radio frequency signal, process the radio frequency signal accordingly, and then sample to obtain a digital sampling sequence.

[0066] More specifically, in step S1 above, after receiving the radio frequency signal, it undergoes down-conversion and filtering to obtain the corresponding baseband signal, which is then sampled by the ADC analog-to-digital converter at a frequency several times the symbol rate, ultimately yielding a digital sampling sequence.

[0067] In this embodiment, sampling is performed at 20 times the symbol rate.

[0068] Down-conversion shifts high-frequency radio frequency signals to more easily processed frequencies. Filtering removes high-frequency components and noise generated during down-conversion; specifically, bandpass filters can be used for this purpose.

[0069] Step S2: Then, perform cross-correlation operation between the digital sampling sequence and the locally stored synchronization word, and take the maximum value of the found correlation peak as the initial synchronization position.

[0070] Step S2 is the coarse synchronization process for the correlation peak. It performs cross-correlation calculations on the sampled data obtained in step S1 and the locally stored synchronization word. By scanning the entire calculation result, the maximum value point of the correlation peak can be found. The sampling position corresponding to this maximum value point is marked as the initial synchronization position P0, thereby completing frame synchronization and coarse symbol synchronization.

[0071] After completing step S2, proceed to step S3, and obtain the optimal symbol synchronization position near the initial synchronization position through soft decision and SNR calculation.

[0072] In this embodiment, the coarse synchronization position P0 obtained from the coarse synchronization process of the correlation peak in step S2 is not the optimal sampling point. It is necessary to find the optimal sampling point with the highest signal-to-noise ratio near the coarse synchronization position P0.

[0073] More specifically, it mainly includes the following sub-steps, including three processes: determining candidate sampling points, soft decision and SNR calculation, and comparing and selecting the optimal synchronization position.

[0074] Step S31: Select several sampling points near the initial synchronization position as candidate points.

[0075] More specifically, for step S31 above, a candidate sampling point set is determined by selecting the sampling points adjacent to the coarse synchronization position P0 as candidate points.

[0076] In this embodiment, three candidate points are selected: {P0-1, P0, P0+1}.

[0077] Step S32: Perform soft decision and SNR calculation based on the above candidate points.

[0078] For each candidate position, the soft decision and SNR calculation process are performed sequentially.

[0079] In this embodiment, the candidate position P_candidate is used as an example for explanation.

[0080] The soft decision process includes the following steps: starting from a candidate position, demodulate the entire current data frame to obtain the soft information sequence for each symbol.

[0081] In this embodiment, P_candidate is used as the symbol starting point, and the entire current data frame (containing known synchronization words and unknown data payloads) is demodulated to obtain the soft signal sequence X_s for each symbol, which is represented as [xs_1, xs_2, ..., xs_N], where N is the frame length; where the soft information can be the amplitude, phase or likelihood ratio information output by the demodulator.

[0082] SNR calculation mainly includes the following process.

[0083] Step S321: Generate an ideal reference signal sequence, perform a hard decision on each value xs_i in the soft information sequence, and map it to the closest standard modulation symbol value.

[0084] Specifically, in order to estimate the noise, an ideal reference signal sequence S_s needs to be generated, which is represented as [ss_1, ss_2, ..., ss_N], where N is the frame length.

[0085] Then, a hard decision is made on each value xs_i in the soft signal sequence X_s, mapping it to the nearest standard modulation symbol value.

[0086] In this embodiment, for 4FSK, the standard values ​​are +3, +1, -1, -3.

[0087] Step S322: Calculate the average value of the mapping results respectively, calculate the scaling factor k based on the average value, and obtain the sequence S_s by multiplying the scaling factor k and the standard modulation symbol value.

[0088] More specifically, the xs_i with hard decision results of +3 and -3 are extracted respectively, and their average values ​​mean_3 and mean_-3 are calculated. The xs_i with hard decision results of +1 and -1 are extracted respectively, and their average values ​​mean_1 and mean_-1 are calculated.

[0089] The scaling factor k is equal to the sum of mean_3 and mean_1 minus the sum of mean_-1 and mean_-3, and then divided by 8; this scaling factor reflects the combined effect of channel gain and distortion on signal amplitude.

[0090] The reference sequence is then generated by multiplying the standard symbol value (+3, +1, -1, -3) to be decided by the scaling factor k to obtain each value ss_i in the reference information sequence S_s, that is, ss_i is equal to the hard decision result multiplied by the scaling factor k; this reference signal sequence S_s can be regarded as the soft information that should be obtained "ideally" after passing through the same channel distortion.

[0091] Step S323: Calculate the total energy and noise energy, obtain the signal energy based on the total energy and noise energy, and finally estimate the signal-to-noise ratio.

[0092] The calculation process of signal-to-noise ratio (SNR) mainly includes the total energy calculation process, noise energy calculation process, signal energy calculation process, and SNR estimation process.

[0093] For the calculation of total energy: the total energy E_total is equal to the sum of the squares of the absolute values ​​of the signal samples xs_i; where i ranges from 1 to N.

[0094] For noise energy calculation: Noise energy E_noise is equal to the sum of the squares of the absolute values ​​of the differences between xs_i and ss_i, where i ranges from 1 to N. This calculation quantifies the difference between the received signal and the ideal reference signal, i.e., noise and distortion.

[0095] For signal energy calculation: signal energy E_signal is the difference between the total energy E_total and the noise energy E_noise.

[0096] Signal-to-noise ratio estimation: The signal-to-noise ratio SNR_est_dB is equal to 10 multiplied by the common logarithm of the ratio of signal energy E_signal to noise energy E_noise.

[0097] Step S33: Select the position corresponding to the candidate point with the largest SNR value as the final optimal symbol synchronization position.

[0098] In more detail, the optimal synchronization positions are compared and selected. The SNR_est_dB values ​​for the three candidate positions P0-1, P0, and P0+1 are calculated respectively. The position corresponding to the candidate point with the largest SNR value is selected as the final optimal symbol synchronization position P_optimal.

[0099] Step S4: Demodulate the data payload at the optimal symbol synchronization position to complete the processing of the current frame.

[0100] At the optimal symbol synchronization position P_optimal, the subsequent data payload is sampled and decided to complete the demodulation of the data.

[0101] Based on this, this embodiment also includes step S5, referring to Figure 1 Step S5 mainly includes maintaining and fine-tuning the synchronization of unsynchronized word frames.

[0102] The synchronization maintenance process for frames without synchronization words includes: maintaining the clock phase corresponding to the optimal synchronization position P_optimal determined in the previous frame as the initial synchronization position P_maintain for the current frame.

[0103] The fine-tuning process for unsynchronized frames includes: selecting candidate points near the initial synchronization position P_maintain of the current frame, calculating the SNR value of each point, and selecting the point with the highest SNR as the new synchronization position after fine-tuning.

[0104] In more detail, the process of fine-tuning the signal-to-noise ratio (SNR) synchronization is repeated. However, since there is no known synchronization word as an absolute reference, when calculating the reference sequence S_s, the soft decision result of the current frame is directly used for hard decision-making, and the scaling factor k is calculated. Similarly, candidate points (such as P_maintain-1, P_maintain, P_maintain+1) are selected near the initial synchronization position P_maintain, the SNR value of each point is calculated, and the point with the highest SNR is selected as the new synchronization position after fine-tuning. This method uses the decision result itself as a reference to achieve fine-tuning under blind synchronization.

[0105] In this technical solution, the synchronization state established in the previous frame is used for maintenance, and the signal-to-noise ratio comparison mechanism is also applied for fine-tuning, so as to achieve long-term stable synchronization even without a synchronization word.

[0106] This embodiment also proposes a symbol synchronization device based on signal-to-noise ratio optimization, referring to... Figure 2 It includes a correlation peak synchronization module and a signal-to-noise ratio (SNR) calculation and selection module connected to the correlation peak synchronization module. The SNR calculation and selection module is also connected to the demodulation module, and the demodulation module is also connected to the synchronization maintenance and fine-tuning module.

[0107] Among them, the correlation peak synchronization module mainly performs the correlation peak coarse synchronization process in step S2.

[0108] The signal-to-noise ratio calculation and selection module can execute the fine synchronization process of signal-to-noise ratio in step S3, which is the core step of the present invention.

[0109] The main function of the demodulation module is to demodulate the data at the optimal position.

[0110] The synchronization maintenance and fine-tuning module can perform the synchronization maintenance and fine-tuning process of the asynchronous word frame in step S5 above.

[0111] The signal-to-noise ratio (SNR) calculation and selection module mainly includes a soft information extraction unit, a reference signal generation unit, an SNR calculation unit, and a comparison and selection unit, which are connected in sequence.

[0112] The soft information extraction unit can generate a soft information sequence X_s based on the candidate point location.

[0113] The reference signal generation unit can make a hard decision on X_s and calculate the scaling factor k to generate a reference soft information sequence S_s.

[0114] The SNR calculation unit can estimate the signal-to-noise ratio based on the X_s and S_s sequences according to the calculation formula involved in Example 1.

[0115] The comparison selection unit can compare SNR and output the optimal position P_optimal.

[0116] In this technical solution, the correlation peak synchronization module, the signal-to-noise ratio calculation and selection module, the demodulation module, and the synchronization maintenance and fine-tuning module are connected in sequence to form a symbol synchronization device based on signal-to-noise ratio optimization. This device can effectively overcome the synchronization error caused by noise and multipath effects, and improve the demodulation performance and decoding success rate of the PDT system.

[0117] The symbol synchronization method based on signal-to-noise ratio optimization involved in this embodiment can solve the problems of insufficient accuracy of the synchronization method based on correlation peak and weak synchronization maintenance capability during the absence of synchronization words.

[0118] The method in this embodiment can further introduce a fine synchronization step based on signal-to-noise ratio comparison after the initial synchronization is completed through the correlation peak. By calculating and comparing the signal-to-noise ratio of the candidate sampling point positions, the optimal symbol synchronization position is selected, thereby significantly improving the synchronization accuracy.

[0119] The method in this embodiment also utilizes the synchronization state established in the previous frame for maintenance, and similarly applies a signal-to-noise ratio comparison mechanism for fine-tuning, thereby achieving long-term stable synchronization even without a synchronization word.

[0120] The settings in this embodiment can effectively overcome synchronization errors caused by noise and multipath effects, and improve the demodulation performance and decoding success rate of the PDT system.

[0121] This embodiment can bring about the following technical effects, including the following three points.

[0122] First, high synchronization accuracy: Unlike the traditional approach that relies solely on the shape of correlation peaks for judgment, this embodiment innovatively uses the signal-to-noise ratio (SNR), an indicator that directly reflects the quality of the communication link, as the final criterion for judging the synchronization position. By selecting the sampling point with the highest SNR, the optimal sampling moment with the strongest anti-interference capability and the largest eye diagram opening can be accurately located, fundamentally improving synchronization accuracy.

[0123] Second, it offers excellent synchronization stability: the method in this embodiment is not only applicable to frames containing synchronization words, but also innovatively applied to the synchronization maintenance and fine-tuning of voice frames without synchronization words. This allows the system to maintain high-precision synchronization throughout the entire communication process, avoiding the gradual loss of synchronization that may occur when relying solely on a phase-locked loop, and greatly enhancing the stability of long-term synchronization.

[0124] Third, it has low implementation complexity and high practicality: the method in this embodiment is easy to implement in existing digital signal processors or FPGAs, and has high engineering application value and market promotion prospects.

Claims

1. A symbol synchronization method based on signal-to-noise ratio optimization, characterized in that, Includes the following steps: S1 receives radio frequency signals, processes them, and then samples them to obtain a digital sampling sequence. S2, perform cross-correlation operation between the digital sampling sequence and the locally stored synchronization word, and take the maximum value of the found correlation peak as the initial synchronization position; S3, through soft decision and SNR calculation, obtains the optimal symbol synchronization position near the initial synchronization position; S4, demodulate the data payload at the optimal symbol synchronization position to complete the processing of the current frame.

2. The symbol synchronization method based on signal-to-noise ratio optimization according to claim 1, characterized in that, Step S3 includes the following steps: S31, Select several sampling points near the initial synchronization position as candidate points; S32, perform soft decision and SNR calculation based on the above candidate points; S33. Select the position corresponding to the candidate point with the largest SNR value as the final optimal symbol synchronization position.

3. A symbol synchronization method based on signal-to-noise ratio optimization according to claim 1 or 2, characterized in that, Step S1 includes: after the radio frequency signal is down-converted and filtered, a baseband signal is obtained, and then sampled by an analog-to-digital converter at a frequency several times the symbol rate to obtain a digital sampling sequence.

4. A symbol synchronization method based on signal-to-noise ratio optimization according to claim 1 or 2, characterized in that, It also includes step S5, which includes maintaining and fine-tuning the synchronization of the unsynchronized word frame. The process of maintaining the synchronization of the unsynchronized word frame includes: using the clock phase corresponding to the optimal synchronization position P_optimal determined in the previous frame as the initial synchronization position P_maintain for the current frame.

5. The symbol synchronization method based on signal-to-noise ratio optimization according to claim 4, characterized in that, The fine-tuning process of the unsynchronized word frame includes: selecting candidate points near the initial synchronization position P_maintain of the current frame, calculating the SNR value of each point, and selecting the point with the highest SNR as the new synchronization position after fine-tuning.

6. A symbol synchronization method based on signal-to-noise ratio optimization according to claim 1 or 2, characterized in that, The soft decision includes: using a candidate position as the symbol starting point, demodulating the entire current data frame to obtain a soft information sequence for each symbol.

7. The symbol synchronization method based on signal-to-noise ratio optimization according to claim 6, characterized in that, The SNR calculation includes: S321, Generate an ideal reference signal sequence, perform hard decision on each value xs_i in the soft information sequence, and map it to the nearest standard modulation symbol value; S322, calculate the average value of the mapping results respectively, calculate the scaling factor k based on the average value, and obtain the sequence S_s by multiplying the scaling factor k and the standard modulation symbol value; S323 calculates the total energy and noise energy, obtains the signal energy based on the total energy and noise energy, and finally estimates the signal-to-noise ratio.

8. The symbol synchronization method based on signal-to-noise ratio optimization according to claim 7, characterized in that, The soft information sequence X_s is represented as [xs_1, xs_2, ..., xs_N], and the ideal reference signal sequence S_s is represented as [ss_1, ss_2, ..., ss_N], where N is the frame length; the closest standard modulation symbol values ​​are +3, +1, -1, -3.

9. A symbol synchronization method based on signal-to-noise ratio optimization according to claim 8, characterized in that, Step S322 includes: Take out the xs_i with hard decision results of +3 and -3 respectively, and calculate their average values ​​mean_3 and mean_-3; Take out the xs_i with hard decision results of +1 and -1 respectively, and calculate their average mean_1 and mean_-1; The scaling factor k is the sum of mean_3 and mean_1 minus the sum of mean_-1 and mean_-3, then divided by 8.

10. A symbol synchronization device based on signal-to-noise ratio optimization, applicable to the symbol synchronization method based on signal-to-noise ratio optimization as described in any one of claims 1-9, characterized in that, It includes a correlation peak synchronization module and a signal-to-noise ratio (SNR) calculation and selection module connected to the correlation peak synchronization module. The SNR calculation and selection module is also connected to a demodulation module, and the demodulation module is also connected to a synchronization maintenance and fine-tuning module.

Images