Signal sequence recovery method and system, computer equipment and storage medium
By calculating the information gain ratio to filter the best sampling point judgment characteristics, the problem of poor symbol synchronization in satellite communication systems is solved, and high-precision and low-overhead timing recovery and clock recovery are achieved.
Patent Information
- Application Number
- CN202411317243.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-20
- Publication Date
- 2025-02-11
AI Technical Summary
In the prior art, it is difficult to effectively determine whether the output of the interpolation filter loop is the optimal sampling point in a satellite communication system, resulting in poor symbol synchronization.
By calculating the information gain ratio of each judgment feature in different scenarios, the best sampling point judgment features with the best timing recovery accuracy, the lowest calculation overhead and the highest robustness are selected, and the best sampling point is determined and nonlinear interpolation is performed to achieve Symbol synchronization and clock recovery.
It effectively solves the problem of poor symbol synchronization, improves the accuracy and stability of timing recovery, reduces the impact of calculation overhead and noise, and is suitable for different satellites' fast mobile reception scenarios.
Smart Images
Figure CN119420401A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of wireless communication technology, and in particular to a signal sequence recovery method, system, computer equipment and storage medium. Background Art
[0002] In satellite communication systems, timing recovery processing technology is usually used at the signal receiving end to compensate for the poor symbol synchronization caused by the error in the transmit and receive clocks. For various interpolation-based timing recovery technologies, it is of great significance to determine whether the output of the interpolation filter loop is the decision feature of the optimal sampling point. Summary of the invention
[0003] In view of the limitations of the existing optimal sampling point decision features, the present application proposes a signal sequence recovery method, system, computer device and storage medium. By calculating the information gain ratio of each decision feature in different scenarios, the decision features with the best timing recovery accuracy, the lowest computational overhead and the highest robustness can be screened out.
[0004] The technical solutions adopted in this application are as follows:
[0005] In a first aspect, the present application proposes a signal sequence recovery method, which is applied to a signal receiving end of a satellite communication system. The signal sequence recovery method comprises:
[0006] Receiving a target signal sequence transmitted by a signal transmitting end in a satellite communication system, constructing an original optimal sampling point set based on symbol information of the target signal sequence, and constructing an optimal sampling point decision feature set based on a preset feature type;
[0007] Calculating the information gain ratio of each decision feature in the optimal sampling point decision feature set based on the original optimal sampling point set, and screening out the optimal sampling point decision feature;
[0008] The optimal sampling point is determined based on the optimal sampling point decision feature, and nonlinear interpolation is performed on the target signal sequence according to the optimal sampling point, thereby realizing symbol synchronization, namely clock recovery, in the satellite communication process.
[0009] Furthermore, the constructing of the optimal sampling point decision feature set based on the preset feature type includes:
[0010] Acquire data corresponding to preset feature types, wherein the preset feature types include a phase difference between a signal optimal sampling moment and a signal receiving end clock, a timing estimation error, a register flip count of a numerically controlled oscillator, and a signal interpolation multiple;
[0011] Construct the optimal sampling point decision feature set based on the corresponding data of the preset feature type.
[0012] Furthermore, the calculating, based on the original optimal sampling point set, the information gain ratio of each decision feature in the optimal sampling point decision feature set includes:
[0013] Dividing the optimal sampling point decision feature set into multiple sub-decision feature sets, and constructing sampling point intersections that simultaneously satisfy the sub-decision feature sets and the original optimal sampling point set one by one;
[0014] Calculating the empirical entropy of the original optimal sampling point set, and calculating the conditional empirical entropy of the original optimal sampling point set according to the intersection of the sampling points;
[0015] Based on the empirical entropy and conditional empirical entropy of the original optimal sampling point set, the information gain ratio of each sampling point intersection with respect to the original optimal sampling point set is calculated.
[0016] Furthermore, the constructing of the original optimal sampling point set based on the symbol information of the target signal sequence includes: constructing the original optimal sampling point set P = {point i}, where point i is the interpolated symbol, i=1,2,L,n, and n is the interpolation multiple at the signal sending end.
[0017] Furthermore, the constructing of the optimal sampling point decision feature set based on the preset feature type includes:
[0018] Construct the optimal sampling point judgment feature set Judge = {phase, error, turn, inter}, where phase = {α k} is the phase difference set between the optimal sampling time of all symbols and the clock of the signal receiving end, α k represents the phase difference between the optimal sampling time of the kth symbol and the clock of the signal receiving end; error = {ε k} represents the estimated timing error set, ε k is the timing error estimated when sampling to the kth symbol; turn = {δ} represents the set of register flip times of NCO, i.e., numerically controlled oscillator, δ = 1, 2, L, ∞; inter = {χ} is the set of interpolation multiples, χ = 1, 2, L, n; where k = 1, 2, L, m, m is the sampling number, and n is the interpolation multiple of the signal transmitter.
[0019] Furthermore, the calculating, based on the original optimal sampling point set, the information gain ratio of each decision feature in the optimal sampling point decision feature set includes:
[0020] According to the multiple sub-decision feature sets of the optimal sampling point decision feature set, the sampling point intersections that simultaneously satisfy the sub-decision feature set and the original optimal sampling point set are constructed one by one; the sub-decision feature set includes a phase difference set phase = {α k}、Timing error set error={ε k}, register flip count set turn = {δ} and interpolation multiple set inter = {χ};
[0021] Calculating the empirical entropy of the original optimal sampling point set, and calculating the conditional empirical entropy of the original optimal sampling point set according to the intersection of the sampling points;
[0022] Based on the empirical entropy and conditional empirical entropy of the original optimal sampling point set, the information gain ratio of each sampling point intersection with respect to the original optimal sampling point set is calculated.
[0023] Furthermore, the step of constructing the intersection of sampling points that simultaneously satisfy the sub-decision feature set and the original optimal sampling point set includes:
[0024] Based on the phase difference set phase = {α k}, the original optimal sampling point set P = {point i} is divided into m subsets, denoted by C = {C1, C2, L, C m}; will exist in the original optimal sampling point set P = {point i} and C k The set of sampling points in is denoted as CP, that is, CP = {CP ki =C k ∩point i}; where k = 1, 2, L, m, and m is the number of sampling times;
[0025] Based on the timing error set error = {ε k}, the original optimal sampling point set P = {point i} is divided into m subsets, denoted as D = {D1, D2, L, D m}; will exist in the original optimal sampling point set P = {point i} and D k The sampling point set in is denoted as DP, that is, DP = {DP ki =D k ∩point i}; where k = 1, 2, L, m, and m is the number of sampling times;
[0026] Based on the register flip count set turn = {δ}, the original optimal sampling point set P = {point i} is divided into several subsets, denoted as E = {E1, E2, L, E ∞}; will exist in the original optimal sampling point set P = {point i} and the set of sampling points in Ej is denoted as EP, that is, EP = {EP ji =E j ∩point i}; where j = 1, 2, L, ∞;
[0027] Based on the interpolation multiple set inter = {χ}, the original optimal sampling point set P = {point i} is divided into n subsets, denoted as F = {F1, F2, L, F n}; will exist in the original optimal sampling point set P = {point i} and F i The set of sampling points in is denoted as FP, that is, FP = {FP ii =F i ∩point i}; wherein, i=1,2,L,n, and n is the interpolation multiple of the signal sending end.
[0028] Furthermore, the empirical entropy calculation method of the original optimal sampling point set includes:
[0029]
[0030] Among them, H(P) is the original optimal sampling point set P = {point i}'s empirical entropy, |point i | for point i The number of elements contained, |P| is the number of elements contained in P.
[0031] Further, the calculating the conditional empirical entropy of the original optimal sampling point set according to the intersection of the sampling points includes:
[0032] Calculate the conditional empirical entropy H(P|phase), H(P|error), H(P|turn) and H(P|inter) of the original optimal sampling point set P under the constraints of the sampling point sets CP, DP, EP and FP respectively:
[0033]
[0034]
[0035]
[0036]
[0037] Among them, |CPki |、|DP ki |、|EP ji |with|FP ii |To respectively set CP ki , DP ki ,EP ji and FP ii The number of elements contained.
[0038] Furthermore, the calculating of the information gain ratio of the intersection of each sampling point with respect to the original optimal sampling point set based on the empirical entropy and conditional empirical entropy of the original optimal sampling point set includes:
[0039] Calculate the information gain ratio of the sampling point sets CP, DP, EP and FP to the original optimal sampling point set P respectively. R (P, phase), gain R (P,error), gain R (P, turn) and gain R (P,inter):
[0040]
[0041]
[0042]
[0043]
[0044] in,
[0045] Furthermore, the performing nonlinear interpolation on the target signal sequence according to the optimal sampling point includes:
[0046] After determining the optimal sampling point based on the selected optimal sampling point judgment features, the target signal sequence RSS is extracted χ , extraction methods include:
[0047] RSS χ =x(τ+mk+χ:χ:χ*len+τ+mk+χ-1)
[0048] Wherein, x() is the output signal of the analog-to-digital converter in the signal receiving end; mk is the interpolation base point calculated according to the O&M algorithm, k = 1, 2, L, m, m is the number of sampling times; χ = 1, 2, L, n, n is the interpolation multiple of the signal sending end; len is the sequence length after deleting the protection symbol; τ is the number of sampling points of the signal sequence of the signal receiving end obtained by capturing the advance or delay;
[0049] Target signal sequence RSS χ Perform nonlinear interpolation to obtain the timing-compensated signal sequence, denoted as RSS out , the interpolation methods include:
[0050] RSS out =f1×RSS χ +f2×RSS χ +…+f n ×RSS χ
[0051] Among them, f n is the nonlinear coefficient.
[0052] In a second aspect, the present application proposes a signal sequence recovery system, which is applied to a signal receiving end of a satellite communication system, and the signal sequence recovery system comprises:
[0053] A set construction module is configured to receive a target signal sequence transmitted by a signal transmitting end in a satellite communication system, construct an original optimal sampling point set based on symbol information of the target signal sequence, and construct an optimal sampling point decision feature set based on a preset feature type;
[0054] A feature screening module is configured to calculate the information gain ratio of each decision feature in the optimal sampling point decision feature set based on the original optimal sampling point set, and screen out the optimal sampling point decision feature;
[0055] The clock recovery module is configured to determine the best sampling point based on the best sampling point decision feature, and perform nonlinear interpolation on the target signal sequence according to the best sampling point, so as to achieve symbol synchronization, i.e. clock recovery, in the satellite communication process.
[0056] In a third aspect, the present application proposes a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned signal sequence recovery method when executing the computer program.
[0057] In a fourth aspect, the present application proposes a computer-readable storage medium storing a computer program, wherein the computer program implements the above-mentioned signal sequence recovery method when executed by a processor.
[0058] The beneficial effects of this application are:
[0059] 1. Aiming at the problem of poor symbol synchronization at the signal receiving end in the satellite communication system, the present application constructs an original optimal sampling point set and an optimal sampling point decision feature set to derive the information gain ratio of different decision features, and can estimate the degree to which each decision feature can reduce the uncertainty of the optimal sampling point decision, thereby solving the limitations of the existing optimal sampling point decision features; the present application can effectively screen out the optimal decision features with both robustness and accuracy in different satellite fast-moving reception scenarios for precise timing recovery; the present application significantly improves the stability of precise timing while maximizing the reduction of the signal bit error rate and convergence time at the signal receiving end.
[0060] 2. Since the parameter values of various decision features in different satellite communication system reception scenarios are different, this application can be used to evaluate and select the best decision features in different satellite fast-moving reception scenarios to support the subsequent accurate demodulation, decoding and other processes of ground terminal data.
[0061] 3. Compared with the prior art, the capture time of this application is shorter, the jitter affected by noise is smaller, the sensitivity to the difference of transmitted data is lower, and the number of sampling points required for each symbol is far less than 32, the convergence time is shorter, and the timing calculation overhead is smaller. BRIEF DESCRIPTION OF THE DRAWINGS
[0062] Figure 1 This is a flow chart of a signal sequence recovery method according to an embodiment of the present application.
[0063] Figure 2 It is a schematic diagram of timing synchronization at the signal receiving end in a satellite communication system.
[0064] Figure 3 It is a line graph of the bit error rate of QPSK signal in different noise environments. DETAILED DESCRIPTION
[0065] In order to have a clearer understanding of the technical features, purposes and effects of the present application, the specific implementation methods of the present application are now described. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application, that is, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application.
[0066] Wen Shenping et al. from the Naval University of Engineering proposed an interpolation filter design based on the minimum mean square error of the interpolation output signal as the criterion, that is, this technology believes that when the mean square error of the interpolation output signal is the smallest, the signal synchronization at the signal receiving end is the best. Since the interpolation timing recovery technology based on the Gardner algorithm can accurately detect the timing deviation when the difference between adjacent code symbols is large, but when the difference between multiple consecutive code symbols is small, the interpolation timer will find it difficult to estimate the timing deviation, and the synchronization jitter is large at this time. Based on this, Liu Wei et al. from Southwest University of Science and Technology proposed an improved Gardner timing recovery technology, which defines the normalized timing error based on the timing error change rate, and determines the optimal sampling time when the timing error converges to the minimum. Zhou Qinglong et al. from the PLA Information Engineering University pointed out that the timing synchronization technology based on interpolation can use the maximum average power as the criterion for the best sampling point, but at least 32 sampling points are required per symbol to ensure the accuracy of timing compensation.
[0067] In order to solve the problem of poor symbol synchronization at the signal receiving end in satellite communication systems, the timing recovery technology based on the mean square error of the signal has better performance when the signal stability is poor, but it is usually difficult to accurately determine the optimal sampling point when the difference in the transmitted data is small; in the timing recovery technology based on the normalized timing error, the capture time is too long, the jitter is large due to the influence of noise, and the robustness needs to be improved; in the timing recovery technology based on the maximum average power, at least 32 sampling points are required per symbol to ensure the accuracy of timing compensation. A large number of sampling points will sharply increase the calculation overhead of timing. In addition, since the algorithm is sensitive to inter-symbol interference, it will lead to a large bit error rate in actual communication scenarios.
[0068] Based on this, in one embodiment of the present application, a signal sequence recovery method is provided, which can screen out the decision features with the best timing recovery accuracy, the lowest computational overhead and the highest robustness by calculating the information gain ratio of each decision feature in different scenarios. Figure 1 As shown, the signal sequence recovery method of this embodiment includes:
[0069] Receiving a target signal sequence transmitted by a signal transmitting end in a satellite communication system, constructing an original optimal sampling point set based on symbol information of the target signal sequence, and constructing an optimal sampling point decision feature set based on a preset feature type;
[0070] Based on the original optimal sampling point set, the information gain ratio of each decision feature in the optimal sampling point decision feature set is calculated, and the optimal sampling point decision feature is screened out;
[0071] The optimal sampling point is determined based on the optimal sampling point decision feature, and the target signal sequence is nonlinearly interpolated according to the optimal sampling point, thereby achieving symbol synchronization, namely clock recovery, in the satellite communication process.
[0072] It should be noted that the Information Gain Ratio is a classification tree learning method based on information gain. Its idea is to select the optimal attribute based on the information gain rate of the attribute. The information gain ratio is used to measure the magnitude of uncertainty reduction. In layman's terms, the value of a feature is characterized by the size of its information gain rate.
[0073] Specifically, constructing an optimal sampling point decision feature set based on preset feature types can be achieved in the following manner: obtaining data corresponding to the preset feature types, the preset feature types including the phase difference between the optimal sampling moment of the signal and the clock of the signal receiving end, the timing estimation error, the number of register flips of the numerically controlled oscillator, and the signal interpolation multiple; constructing an optimal sampling point decision feature set based on the data corresponding to the preset feature types.
[0074] Specifically, the information gain ratio of each decision feature in the optimal sampling point decision feature set is calculated based on the original optimal sampling point set, which can be achieved in the following manner: the optimal sampling point decision feature set is divided into multiple sub-decision feature sets, and the intersection of sampling points that simultaneously satisfies the sub-decision feature set and the original optimal sampling point set is constructed one by one; the empirical entropy of the original optimal sampling point set is calculated, and the conditional empirical entropy of the original optimal sampling point set is calculated based on the empirical entropy and conditional empirical entropy of the original optimal sampling point set; based on the empirical entropy and conditional empirical entropy of the original optimal sampling point set, the information gain ratio of each sampling point intersection with respect to the original optimal sampling point set is calculated.
[0075] More specifically, assuming that the timing recovery process at the signal receiving end in the satellite communication system is as follows Figure 2 As shown. Figure 2 In the example, the input signal x(t) can be a baseband signal after down-conversion, and the symbol period is T i , under the condition of satisfying Nyquist theorem, the signal receiving end is T s As the sampling step, we can get the time domain discrete signal x(mT s ), m is the sampling number, t is the current time; the initial sampling is completed after the signal digital-to-analog conversion and interpolation filtering, and then the timing error detection is performed on the signal output by the initial sampling, and then it is input into the loop filter and the numerically controlled oscillator (NCO) in sequence, and fed back to the interpolation filter for sampling iteration as the output signal after timing synchronization.
[0076] For the interpolation filter at the signal receiving end, assuming that the interpolation multiple at the signal sending end is n, then each symbol after interpolation is composed of n symbols, represented by point1, point2, L, point nSince there is only one optimal sampling point in this scene, there are n possibilities for selecting the optimal sampling point. Based on this, the original optimal sampling point set P = {point i}, where point i is the symbol after interpolation, i=1,2,L,n.
[0077] Specifically, the optimal sampling point decision feature set is constructed based on the preset feature type, which can be implemented in the following ways:
[0078] Construct the optimal sampling point judgment feature set Judge = {phase, error, turn, inter}, where phase = {α k} is the phase difference set between the optimal sampling time of all symbols and the clock of the signal receiving end, α k represents the phase difference between the optimal sampling time of the kth symbol and the clock of the signal receiving end; error = {ε k} represents the estimated timing error set, ε k is the timing error estimated when sampling to the kth symbol; turn = {δ} represents the set of register flip times of NCO, i.e., numerically controlled oscillator, δ = 1, 2, L, ∞; inter = {χ} is the set of interpolation multiples, χ = 1, 2, L, n; where k = 1, 2, L, m, m is the sampling number, and n is the interpolation multiple of the signal transmitter.
[0079] Specifically, the information gain ratio of each decision feature in the best sampling point decision feature set is calculated based on the original best sampling point set, which can be implemented in the following way:
[0080] According to the multiple sub-decision feature sets of the optimal sampling point decision feature set, the sampling point intersections that simultaneously satisfy the sub-decision feature set and the original optimal sampling point set are constructed one by one; the sub-decision feature set includes a phase difference set phase = {α k}、Timing error set error={ε k}, register flip count set turn = {δ} and interpolation multiple set inter = {χ};
[0081] Calculate the empirical entropy of the original optimal sampling point set, and calculate the conditional empirical entropy of the original optimal sampling point set based on the intersection of the sampling points;
[0082] Based on the empirical entropy and conditional empirical entropy of the original optimal sampling point set, the information gain ratio of the intersection of each sampling point with respect to the original optimal sampling point set is calculated.
[0083] More specifically, constructing the intersection of sampling points that simultaneously satisfy the sub-decision feature set and the original optimal sampling point set one by one can be achieved in the following way:
[0084] Based on the phase difference set phase = {α k}, the original optimal sampling point set P = {point i} is divided into m subsets, denoted by C = {C1, C2, L, C m}; will exist in the original optimal sampling point set P = {point i} and C k The set of sampling points in is denoted as CP, that is, CP = {CP ki =C k ∩point i}; where k = 1, 2, L, m, and m is the number of sampling times;
[0085] Based on the timing error set error = {ε k}, the original optimal sampling point set P = {point i} is divided into m subsets, denoted as D = {D1, D2, L, D m}; will exist in the original optimal sampling point set P = {point i} and D k The sampling point set in is denoted as DP, that is, DP = {DP ki =D k ∩point i}; where k = 1, 2, L, m, and m is the number of sampling times;
[0086] Based on the register flip count set turn = {δ}, the original optimal sampling point set P = {point i} is divided into several subsets, denoted as E = {E1, E2, L, E ∞}; will exist in the original optimal sampling point set P = {point i} and the set of sampling points in Ej is denoted as EP, that is, EP = {EP ji =E j ∩point i}; where j = 1, 2, L, ∞;
[0087] Based on the interpolation multiple set inter = {χ}, the original optimal sampling point set P = {point i} is divided into n subsets, denoted as F = {F1, F2, L, F n}; will exist in the original optimal sampling point set P = {point i} and F i The set of sampling points in is denoted as FP, that is, FP = {FP ii =F i ∩point i}; wherein, i=1,2,L,n, and n is the interpolation multiple of the signal sending end.
[0088] Specifically, the empirical entropy of the original optimal sampling point set can be calculated as follows:
[0089]
[0090] Among them, H(P) is the original optimal sampling point set P = {point i}'s empirical entropy, |point i point i The number of elements contained, |P| is the number of elements contained in P.
[0091] Specifically, the conditional empirical entropy of the original optimal sampling point set is calculated according to the intersection of the sampling points, which can be implemented in the following way:
[0092] Calculate the conditional empirical entropy H(P|phase), H(P|error), H(P|turn) and H(P|inter) of the original optimal sampling point set P under the constraints of the sampling point sets CP, DP, EP and FP respectively:
[0093]
[0094]
[0095]
[0096]
[0097] Among them, |CP ki |、|DP ki |、|EP ji |with|FP ii |To respectively set CP ki , DP ki ,EP ji and FP ii The number of elements contained.
[0098] Specifically, based on the empirical entropy and conditional empirical entropy of the original optimal sampling point set, the information gain ratio of the intersection of each sampling point with respect to the original optimal sampling point set is calculated, which can be achieved in the following way:
[0099] Calculate the information gain ratio of the sampling point sets CP, DP, EP and FP to the original optimal sampling point set P respectively. R (P, phase), gain R (P,error), gain R (P, turn) and gain R(P,inter):
[0100]
[0101]
[0102]
[0103]
[0104] in,
[0105] Specifically, nonlinear interpolation of the target signal sequence according to the optimal sampling point can be achieved in the following way:
[0106] After determining the best sampling point based on the selected best sampling point judgment features, the target target signal sequence RSS is extracted χ , the extraction method can be:
[0107] RSS χ =x(τ+mk+χ:χ:χ*len+τ+mk+χ-1)
[0108] Wherein, x() is the output signal of the analog-to-digital converter in the signal receiving end; mk is the interpolation base point calculated according to the O&M algorithm, k = 1, 2, L, m, m is the number of sampling times; χ = 1, 2, L, n, n is the interpolation multiple of the signal sending end; len is the sequence length after deleting the protection symbol; τ is the number of sampling points of the signal sequence of the signal receiving end obtained by capturing the advance or delay;
[0109] Target signal sequence RSS χ Perform nonlinear interpolation to obtain the timing-compensated signal sequence, denoted as RSS out , the interpolation method can be:
[0110] RSS out =f1×RSS χ +f2×RSS χ +…+f n ×RSS χ
[0111] Among them, f n is the nonlinear coefficient.
[0112] In another embodiment of the present application, the signal sequence recovery method is implemented by the following steps:
[0113] 1) At the signal receiving end of the satellite communication system, four decision features are selected to construct the optimal sampling point decision feature set, where the four decision features are: the phase difference between the optimal sampling moment of the signal and the clock of the signal receiving end, the timing estimation error, the number of NCO register flips, and the signal interpolation multiple.
[0114] 2) Determine the intersection of sampling points:
[0115] Based on the decision feature 1, i.e., the phase difference between the optimal sampling moment of the signal and the clock of the signal receiving end, construct the intersection of the original optimal sampling point set P and the optimal sampling point subset obtained by dividing according to the decision feature 1;
[0116] Based on the second decision feature, i.e., the timing estimation error, the intersection of the original optimal sampling point set P and the optimal sampling point subset obtained by dividing according to the second decision feature is constructed;
[0117] Based on the decision feature three, i.e., the number of NCO register flips, construct the intersection of the original optimal sampling point set P and the optimal sampling point subset obtained by dividing according to the decision feature three;
[0118] Based on decision feature four, ie, the signal interpolation multiple, the intersection of the original optimal sampling point set P and the optimal sampling point subset obtained by dividing according to decision feature four is constructed.
[0119] 3) Calculate the information gain ratio of each decision feature:
[0120] According to the formula Calculate the empirical entropy H(P) of the original optimal sampling point set P;
[0121] Calculate the conditional empirical entropy H(P|phase), H(P|error), H(P|turn) and H(P|inter) respectively;
[0122] Calculate the information gain ratio gain respectively R (P, phase), gain R (P,error), gain R (P, turn) and gain R (P,inter).
[0123] 4) According to the information gain ratio of each decision feature with respect to the original optimal sampling point set P, the one with the maximum value is selected as the optimal sampling point decision feature, and the output signal RSS of the corresponding analog-to-digital converter is extracted after obtaining the optimal sampling point. χ .
[0124] 5) Output signal RSS χ By performing nonlinear interpolation, the signal sequence after precise timing compensation can be obtained.
[0125] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity of description, they are expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of the actions described, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.
[0126] In the signal receiving end of the current satellite communication system, in order to verify the correctness of the signal sequence recovery method of the present application, the signal transmitting end is set to use QPSK modulation in the computing system, the bit rate is 5Mb / s, the roll-off factors of the root raised cosine shaping filter and the matched filter are both α=0.6, the signal transmitting end uses 5 times interpolation, the loop coefficient is 0.1, and the nonlinear interpolation coefficients f1, f2, L, and f5 are respectively taken as Based on the above conditions, the calculation results are shown in Table 1. d is the fractional part of the interval between the optimal sampling time and the base point, which can characterize the robustness and effectiveness of the synchronized algorithm. From the results in Table 1, we can see that Δ d Jitter will occur at the beginning, and then gradually converge to a stable value. Since the earlier the convergence and the smaller the stable value, the better the synchronization performance of the corresponding algorithm, the convergence time of the signal sequence recovery method of the present application is 0.0029s, which is shorter than the other two algorithms, and Δ d The mean square error of the two algorithms is smaller than that of the other two algorithms, which indicates the noise resistance and accuracy of the signal sequence recovery method of the present application.
[0127] Table 1Δ d The mean square error
[0128]
[0129] like Figure 3 The figure shows the bit error rate of the QPSK signal in different noise environments. Specifically, the signal sequence recovery method of the present application, the Garden interpolation method based on normalized timing error, and the Garden interpolation method are used as examples to show the change in the bit error rate of the QPSK signal when the simulated signal is 186 symbols under the Gaussian channel. Figure 3 It can be seen that the signal sequence recovery method of the present application has a significant performance improvement when the signal-to-noise ratio is greater than or equal to 6dB; when the signal-to-noise ratio is 18dB, the bit error rate is reduced by about 2 orders of magnitude compared with the other two algorithms.
[0130] In summary, compared with the shortcomings of the prior art, in the signal receiving end of the current satellite communication system, the signal sequence recovery method of the present application has a shorter capture time, less jitter affected by noise, less sensitivity to the differences in transmitted data, and the number of sampling points required for each symbol is far less than 32, the convergence time is short, and the timing calculation overhead is small.
[0131] In another embodiment of the present application, a signal sequence recovery system is provided, which is applied to a signal receiving end of a satellite communication system. The signal sequence recovery system includes a set construction module, a feature screening module and a clock recovery module, wherein:
[0132] The set construction module is configured to receive a target signal sequence transmitted by a signal transmitter in a satellite communication system, construct an original optimal sampling point set based on symbol information of the target signal sequence, and construct an optimal sampling point decision feature set based on a preset feature type.
[0133] The feature screening module is configured to calculate the information gain ratio of each decision feature in the optimal sampling point decision feature set based on the original optimal sampling point set, and screen out the optimal sampling point decision feature.
[0134] The clock recovery module is configured to determine the best sampling point based on the best sampling point decision feature, and perform nonlinear interpolation on the target signal sequence according to the best sampling point, so as to achieve symbol synchronization, i.e. clock recovery, in the satellite communication process.
[0135] In another embodiment of the present application, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the signal sequence recovery method of embodiment 1 is implemented. The computer program may be in source code form, object code form, executable file, or some intermediate form.
[0136] In another embodiment of the present application, a computer-readable storage medium is provided, which stores a computer program, and the computer program, when executed by a processor, implements the signal sequence recovery method of Example 1. The computer program may be in source code form, object code form, executable file, or some intermediate form, etc. The storage medium includes: any entity or device capable of carrying computer program code, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content contained in the storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the storage medium does not include electric carrier signals and telecommunication signals.
[0137] The above is only a preferred embodiment of the present application. It should be understood that the present application is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge in the relevant field. The changes and modifications made by those skilled in the art shall not deviate from the spirit and scope of the present application, and shall be within the scope of protection of the claims attached to the present application.
Claims
1. A signal sequence recovery method, applied to a signal receiving end of a satellite communication system, characterized in that: The signal sequence recovery method comprises: Receiving a target signal sequence transmitted by a signal transmitting end in a satellite communication system, constructing an original optimal sampling point set based on symbol information of the target signal sequence, and constructing an optimal sampling point decision feature set based on a preset feature type; Calculating the information gain ratio of each decision feature in the optimal sampling point decision feature set based on the original optimal sampling point set, and screening out the optimal sampling point decision feature; The optimal sampling point is determined based on the optimal sampling point decision feature, and nonlinear interpolation is performed on the target signal sequence according to the optimal sampling point, thereby realizing symbol synchronization, namely clock recovery, in the satellite communication process.
2. A signal sequence recovery method according to claim 1, characterized in that: The step of constructing an optimal sampling point decision feature set based on a preset feature type includes: Acquire data corresponding to preset feature types, wherein the preset feature types include a phase difference between a signal optimal sampling moment and a signal receiving end clock, a timing estimation error, a register flip count of a numerically controlled oscillator, and a signal interpolation multiple; Construct the optimal sampling point decision feature set based on the corresponding data of the preset feature type.
3. A signal sequence recovery method according to claim 1, characterized in that: Calculating the information gain ratio of each decision feature in the optimal sampling point decision feature set based on the original optimal sampling point set includes: Dividing the optimal sampling point decision feature set into multiple sub-decision feature sets, and constructing sampling point intersections that simultaneously satisfy the sub-decision feature sets and the original optimal sampling point set one by one; Calculating the empirical entropy of the original optimal sampling point set, and calculating the conditional empirical entropy of the original optimal sampling point set according to the intersection of the sampling points; Based on the empirical entropy and conditional empirical entropy of the original optimal sampling point set, the information gain ratio of each sampling point intersection with respect to the original optimal sampling point set is calculated.
4. A signal sequence recovery method according to claim 1, characterized in that: The method of constructing an original optimal sampling point set based on the symbol information of the target signal sequence includes: constructing an original optimal sampling point set P = {point i }, where point i is the interpolated symbol, i=1,2,L,n, and n is the interpolation multiple at the signal sending end.
5. A signal sequence recovery method according to claim 4, characterized in that: The step of constructing an optimal sampling point decision feature set based on a preset feature type includes: Construct the optimal sampling point judgment feature set Judge = {phase, error, turn, inter}, where phase = {α k } is the phase difference set between the optimal sampling time of all symbols and the clock of the signal receiving end, α k represents the phase difference between the optimal sampling time of the kth symbol and the clock of the signal receiving end; error = {ε k } represents the estimated timing error set, ε k is the timing error estimated when sampling to the kth symbol; turn = {δ} represents the set of register flip times of NCO, i.e., numerically controlled oscillator, δ = 1, 2, L, ∞; inter = {χ} is the set of interpolation multiples, χ = 1, 2, L, n; where k = 1, 2, L, m, m is the sampling number, and n is the interpolation multiple of the signal transmitter.
6. A signal sequence recovery method according to claim 5, characterized in that: The calculating, based on the original optimal sampling point set, the information gain ratio of each decision feature in the optimal sampling point decision feature set comprises: According to the multiple sub-decision feature sets of the optimal sampling point decision feature set, the sampling point intersections that simultaneously satisfy the sub-decision feature set and the original optimal sampling point set are constructed one by one; the sub-decision feature set includes a phase difference set phase = {α k }、Timing error set error={ε k }, register flip count set turn = {δ} and interpolation multiple set inter = {χ}; Calculating the empirical entropy of the original optimal sampling point set, and calculating the conditional empirical entropy of the original optimal sampling point set according to the intersection of the sampling points; Based on the empirical entropy and conditional empirical entropy of the original optimal sampling point set, the information gain ratio of each sampling point intersection with respect to the original optimal sampling point set is calculated.
7. A signal sequence recovery method according to claim 6, characterized in that: The step of constructing the intersection of sampling points that simultaneously satisfy the sub-decision feature set and the original optimal sampling point set comprises: Based on the phase difference set phase = {α k }, the original optimal sampling point set P = {point i } is divided into m subsets, denoted by C = {C1, C2, L, C m }; will exist in the original optimal sampling point set P = {point i } and C k The set of sampling points in is denoted as CP, that is, CP = {CP ki =C k ∩point i }; where k = 1, 2, L, m, and m is the number of sampling times; Based on the timing error set error = {ε k }, the original optimal sampling point set P = {point i } is divided into m subsets, denoted as D = {D1, D2, L, D m }; will exist in the original optimal sampling point set P = {point i } and D k The sampling point set in is denoted as DP, that is, DP = {DP ki =D k ∩point i }; where k = 1, 2, L, m, and m is the number of sampling times; Based on the register flip count set turn = {δ}, the original optimal sampling point set P = {point i } is divided into several subsets, denoted as E = {E1, E2, L, E ∞ }; will exist in the original optimal sampling point set P = {point i } and the set of sampling points in Ej is denoted as EP, that is, EP = {EP ji =E j ∩point i }; where j = 1, 2, L, ∞; Based on the interpolation multiple set inter = {χ}, the original optimal sampling point set P = {point i } is divided into n subsets, denoted as F = {F1, F2, L, F n }; will exist in the original optimal sampling point set P = {point i } and F i The set of sampling points in is denoted as FP, that is, FP = {FP ii =F i ∩point i }; wherein, i=1,2,L,n, and n is the interpolation multiple of the signal sending end.
8. A signal sequence recovery method according to claim 7, characterized in that: The empirical entropy calculation method of the original optimal sampling point set includes: Among them, H(P) is the original optimal sampling point set P = {point i }'s empirical entropy, |point i point i The number of elements contained, |P| is the number of elements contained in P.
9. A signal sequence recovery method according to claim 8, characterized in that: The step of calculating the conditional empirical entropy of the original optimal sampling point set according to the intersection of the sampling points includes: Calculate the conditional empirical entropy H(P|phase), H(P|error), H(P|turn) and H(P|inter) of the original optimal sampling point set P under the constraints of the sampling point sets CP, DP, EP and FP respectively: Among them, |CP ki |、|DP ki |、|EP ji |with|FP ii |To respectively set CP ki , DP ki ,EP ji and FP ii The number of elements contained.
10. A signal sequence recovery method according to claim 9, characterized in that: The calculating the information gain ratio of the intersection of each sampling point with respect to the original optimal sampling point set based on the empirical entropy and the conditional empirical entropy of the original optimal sampling point set comprises: Calculate the information gain ratio of the sampling point sets CP, DP, EP and FP to the original optimal sampling point set P respectively. R (P, phase), gain R (P,error), gain R (P, turn) and gain R (P,inter): in, 11. A signal sequence recovery method according to any one of claims 1 to 10, characterized in that: The nonlinear interpolation of the target signal sequence according to the optimal sampling point comprises: After determining the optimal sampling point based on the selected optimal sampling point judgment features, the target signal sequence RSS is extracted χ , extraction methods include: RSS χ =x(τ+mk+χ:χ:χ*len+τ+mk+χ-1) Where x( ) is the output signal of the analog-to-digital converter at the signal receiving end; mk is the interpolation base point calculated according to the O&M algorithm, k = 1, 2, L, m, m is the sampling number; χ = 1, 2, L, n, n is the interpolation multiple at the signal sending end; len is the sequence length after deleting the protection symbol; τ is the number of sampling points of the signal sequence at the signal receiving end obtained by capturing the advance or delay; Target signal sequence RSS χ Perform nonlinear interpolation to obtain the timing-compensated signal sequence, denoted as RSS out , the interpolation methods include: RSS out =f1×RSS χ +f2×RSS χ +…+f n ×RSS χ Among them, f n is the nonlinear coefficient.
12. A signal sequence recovery system, applied to a signal receiving end of a satellite communication system, characterized in that: The signal sequence recovery system comprises: A set construction module is configured to receive a target signal sequence transmitted by a signal transmitting end in a satellite communication system, construct an original optimal sampling point set based on symbol information of the target signal sequence, and construct an optimal sampling point decision feature set based on a preset feature type; A feature screening module is configured to calculate the information gain ratio of each decision feature in the optimal sampling point decision feature set based on the original optimal sampling point set, and screen out the optimal sampling point decision feature; The clock recovery module is configured to determine the best sampling point based on the best sampling point decision feature, and perform nonlinear interpolation on the target signal sequence according to the best sampling point, so as to achieve symbol synchronization, i.e. clock recovery, in the satellite communication process.
13. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the signal sequence recovery method according to any one of claims 1 to 10 is implemented.
14. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the signal sequence recovery method according to any one of claims 1 to 10 is implemented.