A method for modulating and identifying continuous phase signals based on the characteristics of a class constellation diagram

Abstract

The present invention discloses a method for modulating and identifying continuous phase signals based on the characteristics of a class constellation diagram. First, the digital signal to be identified enters a digital receiver and first undergoes parameter estimation processing, including center frequency estimation and bandwidth estimation, and the instantaneous frequency information is extracted using these as prior conditions. Secondly, the instantaneous frequency information is energy-normalized, and a class constellation diagram is constructed. The convergence degree of the class constellation diagram near the zero point and the ±1 points is statistically compared with a threshold to divide the signal set. Then, the number of convergence intervals of the class constellation diagram is statistically used as a feature quantity and compared with a threshold to divide the remaining signal set. Finally, for the remaining signal set after division, the symbol rate of the signal is estimated through the fourth-order spectrum, the normalized frequency variance is calculated and compared with a threshold to complete the final signal division. The present invention has the advantages of low complexity, high recognition rate, strong robustness, and still having a high recognition rate under low signal-to-noise ratio and with frequency offset.

Description

Technical Field

[0001] The present invention belongs to the field of modulation recognition, and particularly relates to a method for continuous phase signal modulation recognition based on the characteristics of a class of constellation diagrams. Background Art

[0002] Continuous phase modulation (CPM) belongs to non-linear digital communication modulation methods, and has the characteristics of continuous phase, constant envelope, and compact frequency band, and is widely used in satellite communication systems. However, in actual satellite communication, the frequency domain similarity between different types of continuous phase signals is high, and it is difficult to identify them.

[0003] Peng Huafu et al. in their published paper "Recognition of Continuous Phase Modulation Signals Based on Instantaneous Amplitude Characteristics" can effectively identify continuous phase signals and non-continuous phase signals by statistically analyzing instantaneous amplitude characteristics. However, the instantaneous amplitude characteristics are easily affected by noise, and the recognition performance is low under low signal-to-noise ratios. Dong Xue in her published paper "Automatic Recognition of Modulation Modes Based on Instantaneous Information and Spectrum Characteristics" proposed a method for continuous phase signal recognition based on instantaneous information and spectrum characteristics. However, parameter estimation is easily affected by the signal-to-noise ratio, and the recognition performance decreases under low signal-to-noise ratios. Hou Yuetao in his published paper "Research on Digital Phase Modulation Signal Recognition and Parameter Estimation Technology" adopted a decision tree classification method based on high-order cumulants to achieve the distinction of multiple types of continuous phase signals. However, the performance of this method decreases when parameter estimation such as frequency offset is inaccurate. Lv Ping et al. in their published paper "A Modulation Recognition Algorithm for CPM Signals and PSK-Type Signals" proposed a method for recognizing continuous phase signals and multiple types of PSK signals based on the characteristics of fractal box dimension, but it is unable to accurately identify different types of continuous phase signals.

[0004] The above modulation recognition algorithms do not pay attention to the internal modulation characteristics of continuous phase signals during feature extraction, and the extraction of the phase characteristics and frequency characteristics of continuous phase signals is insufficient. When multiple types of continuous phase signals exist simultaneously, it is difficult to make targeted recognition. Summary of the Invention

[0005] The object of the present invention is to improve the signal recognition ability of a signal set including {Frequency Shift Keying MSK, Continuous Phase Frequency Shift Keying CPM with a single modulation index, CPM with multiple modulation indices, Shaped Offset Quadrature Phase Shift Keying SOQPSK, Shaped Binary Phase Shift Keying SBPSK}, and a continuous phase signal modulation recognition algorithm method based on the characteristics of a class constellation diagram is proposed, which can complete the recognition of the continuous phase signal set {MSK, CPM with a single modulation index, CPM with multiple modulation indices, SOQPSK, SBPSK}. The blind digital receiver is used to extract the phase modulation information of the continuous phase signal, and the instantaneous frequency extracted from the phase modulation information is used to construct the characteristics of the class constellation diagram. The continuous phase signals in the signal set are recognized through the convergence degree of the class constellation diagram, the number of convergence intervals of the class constellation diagram, and the normalized frequency variance. Theoretical analysis and simulation results show that the proposed continuous phase signal modulation recognition algorithm based on the blind digital receiver and the characteristics of the class constellation diagram has good recognition performance compared with the widely used modulation recognition algorithm based on high-order cumulants at low signal-to-noise ratios. When the signal-to-noise ratio is 3 dB, the average recognition rate can reach more than 90%.

[0006] The technical solution adopted by the present invention to solve its technical problems includes the following steps:

[0007] Step 1: The digital signal to be recognized enters the digital receiver and first performs parameter estimation processing, including central frequency estimation and bandwidth estimation. The estimation method uses the Fourier transform to roughly estimate the central frequency and bandwidth of the signal to be recognized. The estimated values of the central frequency and bandwidth are used as prior conditions and input into the blind digital receiver, and the phase modulation information is extracted by using a closed-loop carrier synchronization loop with an arctangent phase discriminator method to extract the instantaneous frequency information.

[0008] Step 2: The instantaneous frequency information is energy-normalized, and a class constellation diagram is constructed. The convergence degree η1 of the class constellation diagram near zero is compared with the threshold TH1, and the signal set formed by the digital signal to be recognized is divided into {MSK, CPM with a single modulation index, CPM with multiple modulation indices} signals and {SOQPSK, SBPSK} signals.

[0009] Step 3: The convergence degree η2 of the class constellation diagram near ±1 is statistically compared with the threshold TH2, and the signal set {MSK, CPM with a single modulation index, CPM with multiple modulation indices} is divided into two categories: MSK and {CPM with a single modulation index, CPM with multiple modulation indices}. Then, the statistical feature quantity C of the number of convergence intervals of the class constellation diagram is statistically compared with the threshold TH3 to identify the CPM signal with a single modulation index and the CPM signal with multiple modulation indices.

[0010] Step 4: For the signal set {SOQPSK, SBPSK}, estimate the symbol rate of the signal through the fourth - power spectrum to eliminate the influence of the symbol rate on the instantaneous frequency level amplitude of SOQPSK and SBPSK signals. Normalize the frequency variance μ and compare it with the threshold TH4 to identify SOQPSK signals and SBPSK signals.

[0011] The specific implementation of Step 1 is as follows:

[0012] 1 - 1. Establish the received signal model as follows:

[0013] The continuous - phase signals studied in the present invention exist in the ultra - short - wave frequency band of 30 MHz - 300 MHz. The set includes {MSK, single - modulation - index CPM, multi - modulation - index CPM, SOQPSK, SBPSK}, a total of 5 modulation types, and all are full - response modulation methods. The general model of the continuous - phase signal is:

[0014]

[0015] In the formula, A(t) is the amplitude variation function of the signal, f c is the carrier frequency of the signal, is the initial phase of the signal, n0(t) is the Gaussian white noise obeying N(0,σ 2 ), is the phase variation function carrying information symbols, t is the current transmission moment, and α is the current transmission symbol. The expression is:

[0016]

[0017] In the formula, h n is the modulation index, α n is the sequence representing information symbols, N is the total data length, n is the current information symbol number, T b is the duration of a single symbol, and q(t) is the phase pulse shaping function.

[0018] The CPM signals studied in the present invention are quaternary continuous - phase signals, and the information sequence α n ∈{±1, ±3}. Among them, the modulation index h of the single - modulation - index CPM signal is 0.5, and the modulation index h of the multi - modulation - index CPM signal is {4 / 16, 8 / 16}.

[0019] The MSK signal can be regarded as a binary continuous - phase signal, and the information sequence α n ∈{±1}, and the modulation index h = 0.5.

[0020] The SOQPSK signal is a special continuous-phase signal, evolved from the OQPSK signal. The phase change of the SOQPSK signal within one symbol period is 0 or ±π / 2, and the transmitted information sequence α n After precoding, it is transformed from a binary sequence into a ternary sequence, α n ∈{-1, 0, 1}, and the modulation index h = 0.5.

[0021] Phase shaping functions of MSK, CPM, and SOQPSK signals:

[0022]

[0023] There is a differential relationship between symbols of the SBPSK signal. Similarly, the binary information sequence α n needs to be precoded. After coding, α n ∈{-1, 0, 1}, and the modulation index h = 1. Different from MSK, CPM, and SOQPSK signals, the width of the phase shaping function of the SBPSK signal is half a symbol period, and the amplitude is 1 / T b .

[0024] 1 - 2. Parameter estimation is performed on the received signal s(n). In this algorithm, there is a closed-loop carrier synchronization in the process of signal information extraction. Therefore, the requirements for the accuracy of center frequency estimation and bandwidth estimation are not high. So, when the signal modulation type is unknown, the power spectrum is obtained by performing a fast Fourier transform on the information sequence s(n), and the position with the highest peak is selected by traversing the entire signal power spectrum as the estimated signal center frequency to roughly estimate the signal carrier frequency. At the same time, the 3dB bandwidth of the received signal is estimated using power spectrum estimation as the bandwidth estimation result.

[0025] 1 - 3. Carrier synchronization extracts the instantaneous frequency information of the continuous-phase signal:

[0026] The parameter estimation values obtained in step 1 - 2 are used as prior conditions and input into the blind digital receiver, and the phase modulation information is extracted using the closed-loop carrier synchronization loop, and the carrier synchronization loop is a Costas loop.

[0027] Compared with non-continuous-phase signals, the instantaneous frequency of continuous-phase signals is time-varying. Therefore, in the phase-locked loop, the arctangent phase discrimination method shown in Equation (4) is used to quickly capture and track the phase of the continuous-phase signal to be identified to extract the instantaneous frequency information.

[0028] u d (t) = arctan(s in_q (t) / s in_i (t)) (4)

[0029] In the formula, s in_i(t) is the input of the in-phase component branch, s in_q (t) is the input of the quadrature component branch, u d (t) is the output of the carrier synchronization phase discrimination.

[0030] When the signal input to the carrier synchronization loop is noise-free, the signal input to the carrier synchronization loop and the local oscillator signal of the receiver and are multiplied, where f NCO is the frequency of the local oscillator signal, represents the initial phase of the local oscillator signal. When the tracking of the synchronization loop tends to steady state, after filtering out the high-frequency components of the signal by the second-order loop filter, the outputs of the in-phase and quadrature component branches are respectively In the formula, is the phase difference between the local oscillator signal and the input signal. When is the case, s out_i (t) ≈ 1 / 2·A(t) is the unipolar information, and s out_q (t) ≈ 0. At this time, the local oscillator signal and the input signal are of the same frequency and in phase, Derive and subtract its own mean value to remove the influence of the original frequency offset, and the instantaneous frequency information of the signals {MSK, monotone modulation index CPM, multi-modulation index CPM, SOQPSK} can be obtained:

[0031]

[0032] The instantaneous frequency information of the SBPSK signal:

[0033]

[0034] Step 2 is specifically implemented as follows:

[0035] 2-1. Energy normalization of instantaneous frequency information to construct a class constellation diagram: It can be seen from formulas (5)-(6) that the amplitude of the instantaneous frequency level of the continuous phase signal is affected by the symbol period T b Influence, and at the same time, accurate signal parameters cannot be known in the actual VHF satellite channel situation. Therefore, when using the characteristics of the class constellation diagram for signal recognition, the influence of T b on the amplitude of the instantaneous frequency level should be eliminated. In actual digital signal transmission, when the synchronization loop tracking is stable, this algorithm performs energy normalization processing on the instantaneous frequency f Inst (n), and the normalized frequency is denoted as:

[0036]

[0037] The normalized instantaneous frequency of the signals {MSK, single modulation index CPM, multiple modulation index CPM, SOQPSK, SBPSK} is as follows:

[0038]

[0039]

[0040]

[0041]

[0042]

[0043] Since the instantaneous frequency information of different signals after normalization has differences in amplitude and number system, the normalized instantaneous frequency information can be used to construct constellation-like diagram features.

[0044] 2-2. The constellation-like diagrams of the signals {SOQPSK, SBPSK} after normalization converge near zero. Therefore, the convergence degree η1 of the constellation-like diagram near zero can be statistically calculated to divide the signal set into {MSK, single modulation index CPM, multiple modulation index CPM} signals and {SOQPSK, SBPSK} signals. The specific calculation formula is as follows:

[0045]

[0046] In the formula, N is the total data length, and N1 is the number of symbols for which the normalized instantaneous frequency information of the signal to be recognized satisfies Set the threshold TH1. If η1≥TH1, it is recognized as {MSK, single modulation index CPM, multiple modulation index CPM} signal; otherwise, it is recognized as {SOQPSK, SBPSK} signal.

[0047] The specific implementation of step 3 is as follows:

[0048] 3-1. Since the instantaneous frequency of the MSK signal has binary jumps, while the instantaneous frequencies of the two types of CPM signals have multiple - base jumps, and the signals have higher convergence degrees near ±1. Therefore, the convergence degree feature η2 of the constellation-like diagram near ±1 can be used to identify the MSK signal and the CPM signal. The specific calculation formula is as follows:

[0049]

[0050] In the formula, N2 is the number of symbols for which the normalized instantaneous frequency information of the signal to be recognized satisfies Set the threshold TH2. If η2≥TH2, it is recognized as the MSK signal; otherwise, it is recognized as {single modulation index CPM, multiple modulation index CPM} signal.

[0051] For the constellation-like diagram of the single modulation index CPM signal, there are four-point convergences, while for the multi-modulation index, there are multiple-point convergences. Therefore, the single modulation index CPM signal and the multi-modulation index CPM signal can be classified and identified by the statistical feature quantity C of the number of convergence intervals in the constellation-like diagram. The specific calculation formula of C is as follows:

[0052]

[0053] In the formula, is the instantaneous frequency information after normalization of the signal to be identified is the number of symbols in the k-th interval. Among them, the intervals are selected as the 4 intervals where there are obvious four-point convergences in the constellation-like diagram of the single modulation CPM signal as known from Eqs. (9)-(10). The statistical intervals are set as (1.3, 1.4), (0.35, 0.45), (-0.45, -0.35), (-1.4, -1.3). Take the threshold H k = 0.018. In the above k-th interval, if then it is considered that there is a convergence in the constellation-like diagram in the current interval, and C k outputs 1. Otherwise, C k outputs 0. C is the sum of the output results of C k in k intervals.

[0054] Set the threshold TH3. If C ≥ TH3, it is identified as a single modulation index CPM signal; otherwise, it is identified as a multi-modulation index CPM signal.

[0055] Step 4 is specifically implemented as follows:

[0056] 4-1. In the actual VHF satellite channel, when the symbol rate of the signal is unknown, if the recognition method of the convergence degree of the constellation-like diagram is adopted, at this time, it is necessary to eliminate the influence of the single symbol duration T b on the amplitude of the instantaneous frequency. Since the instantaneous frequency information of both SOQPSK and SBPSK signals has ternary jumps and zero-point convergences and high similarity. At the same time, if Eq. (7) is used to eliminate the influence of T b on the instantaneous frequency level amplitude of SOQPSK and SBPSK signals, the discrimination of the output results of the two is not large. Therefore, it is difficult to distinguish the two signals using the characteristics of the convergence degree of the constellation-like diagram. Since the fourth-power spectra of SOQPSK signals and SBPSK signals both have obvious spectral lines at the fourth harmonic frequency, and the spectral line interval is the symbol rate. Therefore, the method of the fourth-power spectrum can be used to accurately estimate the symbol rate of SOQPSK signals and SBPSK signals to obtain the estimated value of the single symbol duration

[0057] 4-2. Let the instantaneous frequencies f of the two types of signals Inst(n) Eliminate T b After the influence, the normalized frequency is obtained, and the variance μ of the normalized frequency is calculated as a feature quantity. The specific calculation formula of μ is as follows:

[0058]

[0059] The feature quantity μ of SOQPSK and SBPSK is significantly different. The normalized frequency variance μ is used to identify SOQPSK signals and SBPSK signals. Set the threshold TH4. If μ≥TH4, it is identified as an SBPSK signal; otherwise, it is identified as an SOQPSK signal.

[0060] The beneficial effects of the present invention are as follows:

[0061] 1. The internal synchronization loop of the receiver has a certain frequency offset capture and tracking ability. Compared with the traditional classification model based on the recognition algorithm of statistical signal features, it can obtain more stable instantaneous signal information.

[0062] 2. The method of distinguishing signals according to the class constellation diagram features of different continuous phase signals makes full use of the unique internal modulation phase characteristics of continuous phase signals.

[0063] 3. When the signal-to-noise ratio is higher than 3dB, the recognition rate of the recognition model for MSK, single modulation index CPM, multi-modulation index CPM, SOQPSK, and SBPSK can all reach more than 90%.

[0064] In summary, the present invention has the characteristics of low complexity, high recognition rate, and strong robustness, and still has the advantage of high recognition rate under low signal-to-noise ratio and with frequency offset. Description of the Drawings

[0065] Figure 1 Signal recognition model design;

[0066] Figure 2(a) is the class constellation diagram of the MSK signal;

[0067] Figure 2(b) is the class constellation diagram of the single modulation index CPM signal;

[0068] Figure 2(c) is the class constellation diagram of the multi-modulation index CPM signal;

[0069] Figure 2(d) is the class constellation diagram of the SOQPSK signal;

[0070] Figure 2(e) is the class constellation diagram of the SBPSK signal;

[0071] Figure 3 The curve of the feature quantity η1 of the zero convergence degree of the class constellation diagram changing with the signal-to-noise ratio;

[0072] Figure 4Recognition rate curves of {MSK, CPM} and {SOQPSK, SBPSK} signals under different signal-to-noise ratios;

[0073] Figure 5 Curve of the concentration degree feature quantity η2 of the ±1 points of the class constellation diagram varying with the signal-to-noise ratio;

[0074] Figure 6 Average recognition rate curves of MSK and CPM signals under different signal-to-noise ratios;

[0075] Figure 7 Curve of the number statistic C of the convergence intervals of the class constellation diagram varying with the signal-to-noise ratio;

[0076] Figure 8 Recognition rate curves of single modulation index CPM and multi-modulation index CPM signals under different signal-to-noise ratios;

[0077] Figure 9 Curve of the instantaneous frequency variance feature quantity μ varying with the signal-to-noise ratio;

[0078] Figure 10 Recognition rate curves of SOQPSK and SBPSK signals under different signal-to-noise ratios;

[0079] Figure 11 Average recognition rate curves of two algorithms under different signal-to-noise ratios. Specific implementation manner

[0080] In order to make the purpose, technical solutions and advantages of this application clearer, the following further details this application in combination with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not used to limit this application. The technical solutions adopted by the present invention:

[0081] Step 1. The complete continuous phase signal modulation recognition algorithm model based on the class constellation diagram features is as Figure 1 shown. The signal to be recognized is first subjected to parameter estimation processing, including center frequency estimation and bandwidth estimation.

[0082] Step 2. Input the parameter estimation values as prior conditions into the blind digital receiver, and use the carrier synchronization loop to extract the phase modulation information so as to extract the instantaneous frequency information. Use the instantaneous frequency information to construct a class constellation diagram as shown in Figure 2. Figure 1 Statistically analyze the zero-point concentration degree feature quantity η1 of the class constellation diagram, compare η1 with the threshold TH1, and divide the signal set into {MSK, single modulation index CPM, multi-modulation index CPM} signals and {SOQPSK, SBPSK} signals.

[0083] Step 3. Figure 1Statistically calculate the concentration degree η2 of the class constellation diagram near the ±1 points, compare it with the threshold TH2, and divide the signal set {MSK, single modulation index CPM, multi-modulation index CPM} into two categories: MSK and {single modulation index CPM, multi-modulation index CPM}. Then statistically calculate the statistical feature quantity C of the number of concentration intervals of the class constellation diagram, compare it with the threshold TH3, and distinguish between single modulation index CPM signals and multi-modulation index CPM signals.

[0084] Step 4: Adopt the method of quartic spectrum to accurately estimate the symbol rate of SOQPSK signals and SBPSK signals, and obtain the estimated value of the single symbol duration Let the instantaneous frequency f of the two types of signals Inst (n) Eliminate T b After the influence, obtain the normalized frequency, and calculate the normalized frequency variance μ as a feature quantity. Figure 1 Among them, statistically calculate the normalized frequency variance μ and compare it with the threshold TH4 to identify SOQPSK signals and SBPSK signals.

[0085] The specific implementation of Step 1 is as follows:

[0086] 1-1. The center frequency and bandwidth estimation adopt power spectrum estimation.

[0087] 1-2. Use the estimated center frequency to down-convert the signal, and use the estimated symbol rate to adjust the loop bandwidth parameter in the carrier synchronization loop.

[0088] The specific implementation of Step 2 is as follows:

[0089] 2-1. Five types of class constellation diagrams constructed using instantaneous frequency information are shown in Figure 2(a), Figure 2(b), Figure 2(c), Figure 2(d), and Figure 2(e).

[0090] 2-2. The zero-point concentration degree feature quantity η1 of the class constellation diagram of the signal to be identified changes with the signal-to-noise ratio as Figure 3 shown. Take the threshold TH1 = 0.225. When η1 >= TH1, it is judged as {SOQPSK, SBPSK} signals, otherwise it is judged as {MSK, single modulation index CPM, multi-modulation index CPM} signals. The recognition rate change curves of {MSK, single modulation index CPM, multi-modulation index CPM} and {SOQPSK, SBPSK} at different signal-to-noise ratios are as Figure 4 shown.

[0091] The specific implementation of Step 3 is as follows:

[0092] 3-1. The ±1 point concentration degree feature quantity η2 of the class constellation diagram of {MSK, single modulation index CPM, multi-modulation index CPM} signals changes with the signal-to-noise ratio as Figure 5As shown, take the threshold TH2 = 0.625. When η2 >= TH2, it is judged as an MSK signal; otherwise, it is judged as a CPM signal. The change curves of the recognition rates of MSK and {monotonic modulation index CPM, multi-modulation index CPM} signals at different signal-to-noise ratios are as Figure 6 shown.

[0093] 3-2. The statistical characteristic quantity C of the number of convergence intervals of the class constellation diagrams of the monotonic modulation CPM signal and the multi-modulation CPM signal changes with the signal-to-noise ratio as Figure 7 shown. Take the threshold TH3 = 3. When C >= TH3, it is judged as a monotonic modulation CPM signal; otherwise, it is judged as a multi-modulation CPM signal. The change curves of the recognition rates of the monotonic modulation CPM signal and the multi-modulation CPM signal at different signal-to-noise ratios are as Figure 8 shown.

[0094] Step 4 is specifically implemented as follows:

[0095] The curve of the normalized frequency variance μ of the SOQPSK signal and the SBPSK signal changing with the signal-to-noise ratio is as Figure 9 shown. Take the threshold TH4 = 4. When μ >= TH4, it is judged as an SBPSK signal; otherwise, it is judged as a SOQPSK signal. The change curves of the recognition rates of the SOQPSK signal and the SBPSK signal at different signal-to-noise ratios are as Figure 10 shown.

[0096] Example:

[0097] The experimental signal set {MSK, monotonic modulation index CPM, multi-modulation index CPM, SOQPSK, SBPSK} is a high signal-to-noise ratio signal in the ultra-short wave satellite channel collected using a satellite antenna provided by a certain research institute of China Electronics Technology Group. The specific parameter settings are as follows: the symbol rate is 1000 Baud / s, the sampling rate is 200 kHz, the carrier frequency is 24 kHz, and the information sequence length N = 400. Add band-limited Gaussian white noise to the collected signals using MATLAB-2021b, and the signal-to-noise ratio is set to 0 - 12 dB.

[0098] Use the algorithm proposed in the present invention and the recognition algorithm in "Research on Digital Phase Modulation Signal Recognition and Parameter Estimation Technology" (Literature 1) to perform recognition simulation tests on the signals under the MATLAB-2021b software. Each test conducts 10,000 Monte Carlo experiments, and counts the number of correct recognitions and calculates the recognition rate. The recognition rates of the two algorithms are as Figure 11 shown.

[0099] From Figure 11It can be seen that, for the continuous-phase signals of the same modulation type within the set under different signal-to-noise ratios, the average recognition rate of the algorithm of the present invention is higher than that of the modulation recognition algorithm based on high-order cumulants proposed in Document 1. At the same time, the algorithm in Document 1 does not have the ability to recognize single-modulation-index CPM and multi-modulation-index CPM signals. Therefore, the overall performance of the algorithm proposed in the present invention is better than that of the algorithm in Document 1. The algorithm of the present invention extracts the constellation-like diagram features of continuous-phase signals through a blind digital receiver with the advantages of anti-noise and anti-frequency offset. The feature extraction is obvious, and it still has good recognition performance for the recognition of multiple types of continuous-phase signals under low signal-to-noise ratios. When the signal-to-noise ratio is higher than 3 dB, the recognition rates for {MSK, single-modulation-index CPM, multi-modulation-index CPM, SOQPSK, SBPSK} signals can all reach more than 90%.

[0100] Finally, it should be noted that the purpose of publishing the embodiments is to help further understand the present invention. However, those skilled in the art can understand that various substitutions and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. Therefore, the present invention should not be limited to the content disclosed in the embodiments, and the scope of protection claimed by the present invention shall be subject to the scope defined by the claims.

Claims

1. A method for modulating and identifying continuous phase signals based on the characteristics of a class constellation diagram, characterized in that It includes the following steps: Step 1: The digital signal to be recognized enters a digital receiver for parameter estimation processing, including center frequency estimation and bandwidth estimation; The estimated values of the center frequency and bandwidth are input into a blind digital receiver as prior conditions, and a closed-loop carrier synchronization loop with an arctangent phase discrimination loop in the phase discrimination part is used to extract phase modulation information and extract instantaneous frequency information; Step 2: Normalize the energy of the instantaneous frequency information and construct a pseudo constellation diagram; Compare the concentration degree η1 of the pseudo constellation diagram near zero with the threshold TH1, and divide the signal set composed of the digital signal to be recognized into {MSK, single modulation index CPM, multi-modulation index CPM} signals and {shaped offset quadrature phase shift keying SOQPSK, shaped binary phase shift keying SBPSK} signals; Step 3: Statistically compare the concentration degree η2 of the pseudo constellation diagram near ±1 with the threshold TH2, and divide the signal set {MSK, single modulation index CPM, multi-modulation index CPM} into two categories: MSK and {single modulation index CPM, multi-modulation index CPM}; Then statistically compare the statistical feature quantity C of the number of convergence intervals of the pseudo constellation diagram with the threshold TH3 to identify single modulation index CPM signals and multi-modulation index CPM signals; Step 4: For the signal set {SOQPSK, SBPSK}, estimate the symbol rate of the signal through the fourth-order spectrum, normalize the frequency variance μ and compare it with the threshold TH4 to identify SOQPSK signals and SBPSK signals.

2. The continuous phase signal modulation recognition method based on the characteristics of the class constellation diagram according to claim 1, characterized in that The specific process of Step 1 is as follows: 1-1. Establish the following received signal model: The digital signal to be recognized is a continuous phase signal existing in the ultra-short wave frequency band of 30 MHz - 300 MHz. The set includes {MSK, single modulation index CPM, multi-modulation index CPM, SOQPSK, SBPSK}, a total of five modulation types, and all are full response modulation methods; The general model of the continuous phase signal is: where, A(t) is the amplitude variation function of the signal, f c is the carrier frequency of the signal, is the initial phase of the signal, n0(t) is Gaussian white noise that follows N(0,σ 2 ), is the phase variation function carrying information symbols, t is the current transmission time, α is the current transmission symbol, and the expression is: where h n is the modulation index, α n is the sequence representing information symbols, N is the total data length, n is the current information symbol number, T b is the duration of a single symbol, and q(t) is the phase pulse shaping function; 1-2. Perform a fast Fourier transform on the information sequence s(n) to obtain the power spectrum, traverse the entire signal power spectrum to select the position with the highest peak as the estimated signal center frequency, roughly estimate the signal carrier frequency, and at the same time use the power spectrum to estimate the 3 dB bandwidth of the received signal as the bandwidth estimation result; 1-3. Carrier synchronization to extract the instantaneous frequency information of the continuous phase signal: Input the parameter estimation values obtained in Step 1-2 into a blind digital receiver as prior conditions, and use a closed-loop carrier synchronization loop to extract phase modulation information. The carrier synchronization loop is a Costas loop; In the phase-locked loop, use the arctangent phase discrimination method shown in the following formula to quickly capture and track the phase of the continuous phase signal to be recognized and extract the instantaneous frequency information: u d (t) = arctan(s in_q (t) / s in_i (t)) where s in_i (t) is the input of the in-phase component branch, s in_q (t) is the input of the quadrature component branch, and u d (t) is the carrier synchronization phase discrimination output; When the signal input to the carrier synchronization loop is noise-free, the signal input to the carrier synchronization loop and the local oscillator signal of the receiver and are multiplied, where f NCO is the frequency of the local oscillator signal, and represents the initial phase of the local oscillator signal; when the tracking of the synchronization loop tends to steady state, after filtering out the high-frequency components of the signal by a second-order loop filter, the in-phase and quadrature component branches output respectively as is the phase difference between the local oscillator signal and the input signal; when is the case, s out_i (t)≈1 / 2·A(t) is a unipolar message, and s out_q (t)≈0. At this time, the local oscillator signal and the input signal are of the same frequency and in phase, Derive and subtract its own mean value to obtain the instantaneous frequency information of the signals {MSK, single modulation index CPM, multi-modulation index CPM, SOQPSK}: The instantaneous frequency information of the SBPSK signal:

3. A method for identifying continuous phase signal modulation based on the characteristics of a class constellation diagram according to claim 2, characterized in that, The specific process of Step 2 is as follows: 2-1. Instantaneous frequency information energy normalization construction type constellation diagram: In actual digital signal transmission, when the synchronization loop tracks stably, perform energy normalization processing on the instantaneous frequency f Inst (n), and at this time, the normalized frequency is denoted as: The normalized instantaneous frequencies of the {MSK, single modulation index CPM, multi-modulation index CPM, SOQPSK, SBPSK} signals are: Use the normalized instantaneous frequency information to construct the pseudo constellation diagram features; 2-2. The convergence degree η1 of the statistical constellation diagram near zero divides the signal set into {MSK, single modulation index CPM, multi-modulation index CPM} signals and {SOQPSK, SBPSK} signals. The specific calculation formula is as follows: where N is the total data length, and N1 is the number of symbols for which the instantaneous frequency information after normalization of the signal to be recognized satisfies ; a threshold TH1 is set. If η1 ≥ TH1, it is recognized as an {MSK, single modulation index CPM, multi-modulation index CPM} signal; otherwise, it is recognized as a {SOQPSK, SBPSK} signal.

4. The continuous phase signal modulation recognition method based on the characteristics of the class constellation diagram according to claim 3, characterized in that The specific process of step 3 is as follows: 3-1. Use the convergence degree feature η2 of the constellation diagram near ±1 to identify MSK signals and CPM signals. The specific calculation formula of η2 is as follows: where N2 is the number of symbols that the instantaneous frequency information after normalization of the signal to be recognized satisfies Set a threshold TH2. If η2 ≥ TH2, it is recognized as an MSK signal; otherwise, it is recognized as a {single modulation index CPM, multiple modulation index CPM} signal. 3-2. Classify and identify single modulation index CPM signals and multi-modulation index CPM signals through the statistical feature quantity C of the number of convergence intervals of the constellation diagram. The specific calculation formula of C is as follows: In the formula, is the instantaneous frequency information after normalization of the signal to be recognized is the number of symbols in the k-th interval, taking the threshold H k , if in the k-th interval then it is considered that there is a convergence of the class constellation diagram in the current interval, C k outputs 1, otherwise C k outputs 0; C is the sum of the C k output results in the k intervals; Set the threshold TH3. If C≥TH3, it is identified as a single modulation index CPM signal; otherwise, it is identified as a multi-modulation index CPM signal.

5. A method for modulating and identifying continuous phase signals based on the characteristics of a class constellation diagram according to claim 4, characterized in that, The specific process of step 4 is as follows: 4-1. The symbol rate of SOQPSK signal and SBPSK signal is accurately estimated by using the quartic spectrum method to obtain the estimated value of the duration of a single symbol 4-2. Let the instantaneous frequencies \(f\) Inst (n) eliminate \(T\) b After the influence, the normalized frequency is obtained, and the variance \(\mu\) of the normalized frequency is calculated as a feature quantity. The specific calculation formula of \(\mu\) is as follows: The feature quantity μ of SOQPSK and SBPSK is significantly different. Use the normalized frequency variance μ to identify SOQPSK signals and SBPSK signals. Set the threshold TH4. If μ≥TH4, it is identified as an SBPSK signal; otherwise, it is identified as a SOQPSK signal.

Images