CN105991492A - Frequency hopping (FH) signal identification method - Google Patents

Abstract

The invention provides a frequency hopping signal recognition method, which is used to identify the frequency hopping signal in the spacecraft measurement and control communication from the messy electromagnetic environment spectrum by using smooth pseudo-Wigner distribution at the complex spacecraft detection site, including the following Steps: step 1, divide the input signal to obtain sliding overlapping windows, so as to obtain a smooth pseudo-Wigner distribution; step 2, obtain the frequency-hopping ridge line based on the smooth pseudo-Wigner distribution, and calculate the quotient difference of the frequency-hopping ridge line Step 3, according to the mutation pulse, calculate the distance between each mutation pulse, so as to obtain the estimation of the frequency hopping period; and step 4, perform parameter threshold iteration on the estimation of the frequency hopping period, so as to identify Output frequency hopping signal and fixed frequency signal. Therefore, the present invention can realize feature recognition for weak frequency hopping signals under low signal-to-noise ratio, and high recognition rate can be achieved above -9dB.

Description

Frequency Hopping Signal Identification Method

technical field

The invention belongs to the field of spacecraft measurement and control wireless communication signal detection, and relates to a frequency modulation signal identification method, which is suitable for identifying frequency hopping signals in spacecraft measurement and control communication from messy electromagnetic environment spectrums at complex spacecraft detection sites.

Background technique

At present, the methods of communication signal modulation recognition are mainly divided into decision theory method and statistical pattern recognition method. The decision theory method is based on probability theory and Bayesian theory in hypothesis testing, which requires more known parameters and a larger amount of calculation. The statistical pattern recognition method uses the statistical correlation moment variable of each order of the signal as a feature to identify different modulation modes. In recent years, researchers have done a lot of work on the identification of modulation modes of single-carrier digital signals and analog signals. Signal analysis methods such as wavelet transform, short-time Fourier transform, and high-order cumulants are used in modulation identification, which can be used for modulation identification. Binary Phase Shift Keying (Binary Phase Shift Keying, hereinafter referred to as BPSK), Quadrature Phase Shift Keying (hereinafter referred to as QPSK), Quadrature Amplitude Modulation (Quadrature Amplitude Modulation, hereinafter referred to as QAM) and analog Identification of signals, etc. However, the detection and identification of frequency hopping signals are less studied.

The frequency hopping signal is a typical non-stationary signal, and these parameters can be obtained effectively only by means of time-frequency analysis. Time-frequency analysis (Time-Frequency analysis, hereinafter referred to as TMA), especially Wigner-Ville Distribution (hereinafter referred to as WVD), is a powerful tool for analyzing non-stationary signals. However, for multi-component signals or single-component signals whose frequency varies nonlinearly with time (eg, frequency-hopping signals), WVD suffers from severe cross-term interference. It is found through simulation that the correct result cannot be obtained by directly using the WVD value to estimate. However, the smoothed pseudo WVD (Smoothed Pseudo WVD, hereinafter referred to as SPWVD) obtained by smoothing twice in the time domain and the frequency domain can effectively suppress the cross interference term.

Therefore, there is an urgent need for a solution that can use SPWVD instead of WVD and pseudo WVD (Pseudo WVD, hereinafter referred to as PWVD) to estimate the carrier frequency of the frequency hopping signal, and on the basis of the frequency hopping modulation mechanism, detect and identify the frequency hopping signal based on SPWVD , which can effectively distinguish frequency hopping and fixed frequency signals.

Contents of the invention

In order to solve the problems existing in the prior art, the present invention proposes a scheme. Firstly, analyze the function of SPWVD in frequency-hopping signal detection to suppress cross-term interference, and then calculate the difference quotient and parameter based on the frequency-hopping ridge line. Threshold iteration and so on, SPWVD is applied to parameter estimation and signal recognition of frequency hopping signals. Finally, the algorithm is simulated by computer, and the performance of the algorithm before and after improvement is compared to verify the feasibility and effectiveness of the algorithm.

The invention provides a frequency hopping signal recognition method, which is used to identify the frequency hopping signal in the spacecraft measurement and control communication from the messy electromagnetic environment spectrum by using smooth pseudo-Wigner distribution at the complex spacecraft detection site, including the following Steps: step 1, divide the input signal to obtain sliding overlapping windows, so as to obtain a smooth pseudo-Wigner distribution; step 2, obtain the frequency-hopping ridge line based on the smooth pseudo-Wigner distribution, and calculate the quotient difference of the frequency-hopping ridge line Step 3, according to the mutation pulse, calculate the distance between each mutation pulse, so as to obtain the estimation of the frequency hopping period; and step 4, perform parameter threshold iteration on the estimation of the frequency hopping period, so as to identify Output frequency hopping signal and fixed frequency signal.

The smooth pseudo-Wigner distribution is used to represent the energy distribution density of the input signal in the time-frequency domain. For a frequency hopping signal, in each complete frequency hopping period, the smooth pseudo-Wigner distribution value is large in the middle and small on both sides, and The minimum occurs at the moment of the frequency jump.

Specifically, in step 1, it is performed: dividing the input signal to obtain sliding overlapping windows; and performing smoothing pseudo-Wigner distribution transformation on the sliding overlapping windows, and rearranging to obtain a time-frequency distribution matrix.

Execute in step 2: Calculate the maximum value of the time-frequency distribution matrix on the frequency axis at each moment to obtain the maximum value sequence, thereby obtaining the frequency hopping ridge line; and perform difference quotient operation on the frequency hopping ridge line to obtain all The frequency hopping moment is used to form an abrupt pulse, thereby obtaining the frequency hopping moment.

Execution in Step 3: Calculate the average distance between frequency hopping moments according to the abrupt pulse, so as to obtain an estimate of the frequency hopping period.

Step 4 includes: selecting the estimated threshold of the frequency hopping period by means of parameter threshold iteration; calculating the variance value of the estimated sequence of the frequency hopping period; and comparing the variance value with the predetermined variance threshold value to identify frequency hopping signal.

Preferably, in Step 4, the threshold at least includes: an upper threshold and a lower threshold.

Correspondingly, it is executed in step four: adopt parameter threshold value iteration to select the estimated upper threshold and lower threshold of the frequency hopping period; calculate the variance value of the estimated sequence of the frequency hopping period; compare the variance value with the variance threshold value comparison; and when the variance value is smaller than the variance threshold value, it is determined as a frequency hopping signal, and when the variance value is greater than the variance threshold value, it is determined as a fixed frequency signal.

Therefore, compared with the prior art, the present invention has the following beneficial effects:

(1) The prior art is to detect and identify frequency hopping signals under high SNR, and the present invention proposes a feature recognition scheme for weak signals under low SNR;

(2) Most of the signal modulation pattern recognition methods in the prior art are based on signal statistics and high-order cumulants, and certain cumulant values of signals with different modulation patterns will be different, and the analog and digital Simulations have been carried out on the recognition of various modulation mode signals. When the signal-to-noise ratio is higher than 10dB, a better recognition effect can be obtained. The present invention is aimed at the characteristics of frequency-hopping signals, and can realize frequency-hopping signal recognition at low signal-to-noise ratios. The recognition can achieve a high recognition rate above -9dB.

Description of drawings

Fig. 1 is the schematic diagram of the maximum value sequence under -5dB for the signal-to-noise ratio involved in the specific embodiment of the present invention;

Fig. 2 is a schematic diagram of the frequency hopping cycle variance of the maximum value method involved in a specific embodiment of the present invention;

FIG. 3 is a schematic diagram of a frequency hopping baseline involved in a specific embodiment of the present invention;

Fig. 4 is a schematic diagram of frequency hopping time estimation involved in a specific embodiment of the present invention;

Fig. 5 is a schematic diagram of a comparison result for period estimation involved in a specific embodiment of the present invention;

Fig. 6 shows the periodic estimation variance of the maximum value method involved in the specific embodiment of the present invention;

Fig. 7 shows the periodic estimated variance of the difference quotient method involved in the specific embodiment of the present invention;

Fig. 8 shows the recognition rate of the frequency hopping signal of the difference quotient method involved in the specific embodiment of the present invention;

Fig. 9 shows the recognition rate of the frequency hopping signal of the maximum value method involved in the specific embodiment of the present invention; And

FIG. 10 is a flow chart of the frequency hopping signal identification process of the present invention.

detailed description

It should be understood that the method for identifying frequency hopping signals detected by complex spacecraft in the present invention proposes to use SPWVD to perform frequency hopping (Frequency Hopping, hereinafter referred to as FH) signal parameter estimation and identification by analyzing frequency hopping systems and existing parameter estimation algorithms , can effectively distinguish frequency hopping and fixed frequency signals. In the implementation, the performance of frequency hopping signal parameter estimation and identification under low signal-to-noise ratio is effectively improved by using methods such as overlapping sliding windows, frequency hopping ridge difference quotient, and parameter threshold iteration. Under the Matlab platform, the detection and estimation performance of the FH signal is analyzed. The results show that when the signal-to-noise ratio is higher than -9dB, effective parameter estimation and identification can be achieved, and frequency-hopping and fixed-frequency signals can be well distinguished. The performance of the algorithm is obviously better than the existing traditional algorithms.

Therefore, on the basis of the SPWVD spectrum, the present invention uses methods such as overlapping sliding windows, frequency-hopping ridge line difference quotient, parameter threshold iteration, etc. to estimate and identify the parameters of the frequency-hopping signal, which can effectively distinguish between frequency-hopping and fixed-frequency signals . In addition, the difference quotient operation is performed on SPWVD to obtain the jump pulse, and combined with the sliding window and iteration threshold, it is used to estimate the frequency jump period.

Specifically, SPWVD is used for frequency hopping signal parameter estimation and modulation pattern recognition, and frequency hopping signal parameters are estimated by using methods such as overlapping windows, frequency hopping baseline derivation, and parameter threshold iteration. At the same time, according to the characteristics of the frequency hopping signal, it is proposed to use the stability of the frequency fluctuation period to separate the frequency hopping and non-frequency hopping signals. The present invention will be described in detail below with reference to the accompanying drawings 1-10 and specific embodiments, so as to further understand the principles, steps, features and advantages of this solution.

Representation of frequency hopping signal

Frequency hopping is a communication method in which frequency hops over time. It has a set of pseudo-random carrier frequencies controlled by a frequency hopping pattern, and the set of all possible carrier frequencies becomes a frequency hopping set. Signals are transmitted on different channels at different times, and the bandwidth covered by all channels is called the frequency hopping bandwidth. The frequency hopping signal is generally generated by non-linear modulation of a series of pulses using a pseudo-randomly generated frequency shift sequence, as shown in the following formula (1).

$\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; \&Sigma;exp</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&phi;</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; )</span>$

Formula 1)

Among them, p(t) is the basic pulse waveform with time width T _b , {f _n } is the pseudo-randomly generated frequency shift sequence, and {φ _n } is the pseudo-random phase sequence when frequency hopping occurs.

Smoothed Pseudo-Wigner-Vill Distribution

The WVD processed by windowing in the time domain and the frequency domain is called SPWVD, specifically as shown in the following formula (2).

$W_z^{\mathrm{SP}}(t,f)={\&Integral;}_{-\&infin;}^{+\&infin;}{\&Integral;}_{-\&infin;}^{+\&infin;}g\left(u\right)h\left(\mathrm{\&tau;}\right)z(t-u+\frac{\mathrm{\&tau;}}{2})z^{*}(t-u+\frac{\mathrm{\&tau;}}{2})e^{\mathrm{jw\&tau;}}\mathrm{dud\&tau;}$ Formula (2)

Wherein, h(t) and g(t) are window functions of odd length, satisfying h(0)=1 and g(0)=1. To realize smooth pseudo WVD algorithm simulation must discretize the continuous WVD. The definition of discrete pseudo-WVD is shown in the following formula (3)

$W_z^P(\mathrm{no},k)=2\underset{m=-(m-1)/2}{\overset{(m-1)/2}{\mathrm{\&Sigma;}}}h\left(m\right)z(\mathrm{no}+m)z^{*}(\mathrm{no}-m)\exp (-j4\mathrm{\&pi;mk}/m)$ Formula (3)

In formula (3), h(m) is a positive real window function with length M, and when the sampling frequency is f _s , t, f and τ are discrete as n, k, and m (specifically as the following formula (4 ) shown).

$f=\frac{f_{\mathrm{the\; s}}k}{2m},t=\frac{\mathrm{no}}{f_{\mathrm{the\; s}}},\mathrm{\&tau;}=\frac{m}{f_{\mathrm{the\; s}}}$ Formula (4)

It can be seen from formula (3) that the discrete pseudo-Wigner distribution can be calculated by discrete Fourier transform (Discrete FourierTest, hereinafter referred to as DFT). Applying DFT to the M-point function c(n, m), the definition of the function c(n, m) is shown in the following formula (5).

$c(\mathrm{no},m)=\left\{\begin{array}{c}g\left(m\right)z(\mathrm{no}+m)z^{*}(\mathrm{no}-m),(0\&le;m\&le;(m-1)/2)\\ g(m-m)z(\mathrm{no}+m-m)z^{*}(\mathrm{no}-m+m),((m+1)/2\&le;m\&le;(m-1))\end{array}\right.$ Formula (5)

Using the relationship between the analytic signal and the original signal in the frequency domain, the construction of z(n) can be obtained from the DFT of the discrete signal s(n). The specific steps are to first do N-point DFT on s(n) to obtain S(k), and construct Z(k) from S(k). When k=0, Z(k)=X(k), k= 1, 2, ... N/2-1, Z(k)=2X(k), and in other cases, Z(k)=0. Finally, inverse discrete Fourier transform is performed on Z(k), z(n)=IDFT[Z(k)].

Parameter Estimation Method of Frequency Hopping Signal Based on SPWVD

The frequency hopping signal is a non-stationary signal whose frequency varies nonlinearly with time, and the traditional Fourier transform method cannot be used to analyze it. Although wavelet transform and short-time Fourier transform can also provide time-varying features of the signal spectrum, they are basically linear decomposition of the signal. The method of time-frequency analysis can provide better energy concentration and time-frequency resolution for dealing with non-stationary time-varying signals. The SPWVD analyzed above is a powerful tool for analyzing non-stationary signals. It can efficiently analyze the time and frequency aggregation of signal energy without any prior knowledge.

The SPWVD distribution can describe the energy distribution density of the signal in the time-frequency domain, and for the frequency hopping signal in each complete frequency hopping period, the SPWVD value of the signal is large in the middle and small on both sides, and the minimum value appears at the moment of frequency hopping. The usual method uses this characteristic to calculate the time of frequency hopping so as to obtain the estimation of the frequency hopping period.

The frequency hopping period of the frequency hopping signal can be estimated by using the characteristic that the SPWVD distribution of the frequency hopping signal has a minimum value in the frequency hopping, but the estimation effect of this method is not ideal when the signal-to-noise ratio is low. Due to the existence of interference noise, the minimum value of the signal SPWVD distribution maximum sequence does not appear only at the time of frequency hopping. As shown in Figure 1, the sequence of maximum values when the SNR is -5dB makes the estimation effect worse. As shown in Figure 2, it shows that when the signal-to-noise ratio changes from -10dB to 10dB, the maximum value method is used to obtain the estimated value of the frequency hopping period Variance.

In order to obtain a better estimation effect when the signal to noise is relatively low, the new frequency hopping period estimation method proposed by the present invention still utilizes the SPWVD distribution of the frequency hopping signal, and calculates the maximum amplitude at each moment along the frequency axis value, record the position of the frequency point where the maximum value occurs, and thus obtain the frequency hopping ridge line. The difference quotient operation is performed on the frequency hopping ridge line, and a sudden pulse will be formed at all frequency hopping moments.

According to the above abrupt pulses, the distance between each abrupt pulse is calculated, that is, the estimation of the frequency hopping period is obtained. Because only at the frequency jump point will there be a relatively large abrupt pulse, so a better estimation effect can still be obtained in the case of a low signal-to-noise ratio. In the period estimation algorithm, an adaptive threshold method based on iteration is adopted to effectively exclude invalid period estimation values. Methods such as overlapping sliding windows, frequency hopping ridge difference quotient, parameter threshold iteration, etc.

suppose is an estimated value of a series of periods, in order to remove invalid period estimates and increase the accuracy of period estimation, the threshold of estimated values is now selected according to the following formula (6):

$T\min =\frac{1}{2}\mathrm{mean}\left(\stackrel{\mathrm{\&Lambda;}}{\mathrm{Ti}}\right);T\max =2\mathrm{mean}\left(\stackrel{\mathrm{\&Lambda;}}{\mathrm{Ti}}\right)$ Formula (6)

right Estimated values greater than Tmax and less than Tmin are discarded.

Implementation process

The carrier frequency of the frequency hopping signal changes periodically pseudo-randomly with time, and the period estimation of the frequency hopping signal can obtain a relatively stable estimated value, but for the fixed frequency signal, since there is no fixed frequency hopping period, noise and modulation The maximum frequency of the signal amplitude will fluctuate randomly within the signal bandwidth range, so the frequency change period value will be an unstable sequence with large fluctuations. The stability of the estimated value of the frequency variation period can thus be used as a feature for frequency hopping signal identification. As shown in Figure 10, the specific steps of the FM signal identification method of the present invention are as follows:

① Divide a sliding overlapping window on the signal x(t), x _i (t), i=1, 2,..., M;

② Perform SPWVD transformation on each x _i (t), and rearrange to obtain its time-frequency distribution matrix s(f, t).

③ Calculate the maximum value on the frequency axis at each moment for s(f, t), obtain the maximum value sequence m(t), and obtain the frequency hopping ridge;

④ Perform difference quotient calculation on the frequency hopping ridge line, and obtain all frequency hopping moments to form abrupt pulses, and further obtain frequency hopping moments;

⑤ Find the average distance between the hopping moments, which is the frequency hopping period estimate;

⑥ parameter threshold iteration, select the threshold of the estimated value: (7), where, for The estimated values greater than Tmax and less than Tmin are discarded;

⑦ Calculate the variance value σ1 of the period estimation sequence; and

⑧ Compare σ1 with the threshold value σ_TH. If σ1<σ_TH, it is determined as a frequency-hopping signal; if σ1>σ_TH, it is determined as a fixed-frequency signal, and then the process ends.

According to the process ①～③, the signal x(t) is divided into sliding overlapping windows, and SPWVD transformation is performed, and then rearranged to obtain its time-frequency distribution matrix. Next, according to the maximum value of the frequency axis, the maximum value sequence m(t) is obtained, thereby obtaining the frequency hopping ridge (as shown in FIG. 3 ). The difference quotient calculation is performed on the frequency hopping ridge line, and a sudden change pulse will be formed at all frequency hopping moments (as shown in Figure 4).

According to the above implementation process ④~⑥, the frequency hopping period can be estimated. Fig. 5 is the frequency hopping period estimation curve obtained by two methods when the signal-to-noise ratio changes from -10dB to 10dB.

It can be seen that the ridge difference quotient method can still obtain relatively stable period estimation when the signal-to-noise ratio is lower than -5dB, but the maximum value method can only obtain stable period estimation when the signal-to-noise ratio is above 0dB.

In the simulation experiment, the frequency hopping set of the frequency hopping signal is 10 carrier frequencies distributed at equal intervals from 1kHz to 10kHz, and the frequency hopping pattern is controlled by a pseudo-random code, including 5 frequency hopping periods. The sampling frequency f _s is set to 2MHz, the frequency hopping frequency f _h is set to 100Hz, that is, the frequency hopping period is 0.01s, and the simulation time T is 0.25s.

According to the above implementation process ⑦, the variance of the frequency hopping signal estimation period can be estimated. Figure 6 shows the estimated periodic variance of fixed-frequency and frequency-hopping signals obtained by the maximum value method, and Figure 7 shows the estimated periodic variance of fixed-frequency and frequency-hopping signals obtained by the difference quotient method.

According to the implementation process ⑧ above, the frequency-hopping signal and the fixed-frequency signal can also be separated according to the estimated period variance. As shown in Figure 6, for the maximum value method, it is difficult to separate the fixed-frequency and frequency-hopping signals when the SNR is below 0dB, while in Figure 7, the fixed-frequency Separate from frequency hopping signals.

In addition, Fig. 8 is a change curve of the recognition rate of the frequency-hopping signal and the fixed-frequency signal when the signal-to-noise ratio changes from -10dB to -5dB in the case of using the difference quotient method. As shown in Figure 8, when the signal-to-noise ratio is higher than -9dB, a higher recognition rate can be obtained for the signal. Figure 9 shows the recognition rates of frequency-hopping signals and fixed-frequency signals obtained by applying the maximum value method. As shown in FIG. 9 , a good recognition rate cannot be obtained when the signal-to-noise ratio is below 5 dB.

It can be seen that the present invention effectively improves the performance of frequency hopping signal parameter estimation under low signal-to-noise ratio, and can accurately estimate the frequency-hopping period under -6dB signal-to-noise ratio. Accurate separation of frequency hopping and non-frequency hopping signals can be achieved above -9dB. In the field of space electronic countermeasures and civilian electromagnetic environment monitoring, especially the development and application of complex spacecraft interference positioning technology, it will have a more important role in promoting and bring better social and economic benefits.

To sum up, the present invention can realize feature recognition for weak frequency hopping signals under low signal-to-noise ratio, and can achieve a high recognition rate above -9dB.

The parts not described in the present invention belong to the known technology in the art.

Claims (8)

1. A frequency hopping signal identification method, used for detecting the scene of complex spacecraft, using smooth pseudo-Wigner distribution, identifying the frequency hopping signal in the spacecraft measurement and control communication from the cluttered electromagnetic environment spectrum, characterized in that, Include the following steps: Step 1, dividing the input signal to obtain a sliding overlapping window, thereby obtaining a smooth pseudo-Wigner distribution; Step 2, obtaining a frequency-hopping ridge line based on the smooth pseudo-Wigner distribution, and performing a quotient difference operation on the frequency-hopping ridge line, thereby obtaining a sudden pulse; Step 3, calculating the distance between each abrupt pulse according to the abrupt pulse, so as to obtain an estimate of the frequency hopping period; and Step 4, perform parameter threshold value iteration on the estimation of the frequency hopping period, so as to identify the frequency hopping signal and the fixed frequency signal. 2. The frequency hopping signal identification method according to claim 1, wherein the smooth pseudo-Wigner distribution is used to represent the energy distribution density of the input signal in the time-frequency domain, For the frequency hopping signal, in each complete frequency hopping period, the values of the smooth pseudo-Wigner distribution are large in the middle and small in both sides, and a minimum value appears at the moment of frequency hopping. 3. the frequency hopping signal identification method according to claim 2, is characterized in that, in described step 1, carries out: dividing the input signal to obtain the sliding overlapping window; and Smooth pseudo-Wigner distribution transformation is performed on the sliding overlapping windows, and rearranged to obtain a time-frequency distribution matrix. 4. the frequency hopping signal identification method according to claim 3, is characterized in that, in described step 2, carries out: Calculate the maximum value on the frequency axis at each moment for the time-frequency distribution matrix to obtain a sequence of maximum values, thereby obtaining the frequency hopping ridge; and A difference quotient operation is performed on the frequency hopping ridge line to obtain all frequency hopping moments to form abrupt pulses, thereby obtaining frequency hopping moments. 5. the frequency hopping signal identification method according to claim 4, is characterized in that, in described step 3, carry out: According to the abrupt pulse, the average distance between the frequency hopping moments is calculated, so as to obtain the estimation of the frequency hopping period. 6. the frequency hopping signal identification method according to claim 5, is characterized in that, described step 4 comprises: Selecting an estimated threshold of the frequency hopping period by means of iteration of the parameter threshold; calculating a variance value of the estimated sequence of frequency hopping periods; and The variance value is compared with a predetermined variance threshold value, thereby identifying the frequency hopping signal. 7. The frequency hopping signal identification method according to claim 6, characterized in that, in the step 4, the threshold at least includes: an upper threshold and a lower threshold. 8. the frequency hopping signal identification method according to claim 7, is characterized in that, in described step 4, carry out: Selecting an estimated upper threshold and a lower threshold of the frequency hopping period by means of parameter threshold iteration; calculating the variance value of the estimated sequence of the frequency hopping period; comparing the variance value to the variance threshold; and When the variance value is smaller than the variance threshold value, it is determined as the frequency hopping signal, and when the variance value is greater than the variance threshold value, it is determined as the fixed frequency signal.