A method for estimating a hop period based on an FFT transform

Abstract

The present application belongs to the field of frequency hopping network station parameter estimation, and particularly relates to a frequency hopping period estimation method based on FFT transformation. In this scenario, mainly containing frequency hopping network station signals and other interference signals, the characteristics of conventional ultrashort wave frequency hopping network station fixed frequency hopping period can be utilized, the time-frequency peak value of frequency hopping pulse is obtained through a time-frequency peak value detection algorithm, the time center sequence of frequency hopping pulse is created according to the time-frequency peak value, and the FFT transformation processing is performed on the time center sequence, so that the peak value coordinate information related to the frequency hopping period can be read in the transformed FFT spectrum, and the value of the frequency hopping period can be calculated according to the information coordinate. The present application can obtain accurate estimation value under the condition of poor detection (more missed detection frequency hopping numbers) and low time resolution of time-frequency diagram, and the method is simple and has good effect.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of frequency hopping communication reconnaissance, and particularly relates to a frequency hopping period estimation method based on FFT transformation. BACKGROUND

[0002] Frequency hopping communication is one of the most commonly used spread spectrum communication modes, and its working principle is that the carrier frequencies of the transmission signals of the transceiver parties are discretely changed according to a predetermined rule, that is, the carrier frequencies used in the communication are randomly changed under the control of a pseudo-random change code. From the implementation mode of the communication technology, frequency hopping communication is a communication mode of multiple frequency shift keying by using a code sequence, and is also a communication system of code-controlled carrier frequency hopping. Frequency hopping communication is widely used in military communication, such as short wave and ultrashort wave radio using frequency hopping technology, and is rapidly permeated into civilian communication, such as mobile communication, data transmission, computer wireless data transmission and wireless local area network.

[0003] Frequency hopping communication reconnaissance usually includes three tasks of frequency hopping signal detection, parameter estimation and signal sorting. Frequency hopping signal detection is to intercept unknown frequency hopping signals of the enemy, including detecting frequency hopping signals mixed in noise; parameter estimation is to estimate the frequency hopping rate, time hopping, frequency hopping pattern, direction of arrival and other parameters of the detected unknown frequency hopping signals; and sorting is to sort the intercepted frequency hopping signals in the network and the network stations in the network by using the estimated parameters, so as to facilitate subsequent demodulation and decryption or tracking interference.

[0004] Frequency hopping signals belong to non-stationary signals, which are composed of multiple time-frequency hopping pulses, and time-frequency analysis is a beneficial tool for analyzing such signals. Therefore, most of the current frequency hopping signal parameter estimation methods are inclined to time-frequency analysis technology, and the premise of part of the frequency hopping parameter estimation is to accurately obtain the time-frequency peak value of the frequency hopping pulse. The frequency hopping period, time hopping, carrier frequency and other parameters are estimated according to the time-frequency peak value of the frequency hopping pulse. The commonly used time-frequency peak value detection algorithms include a fast frequency hopping signal peak value detection algorithm based on two-dimensional second difference, a time-frequency peak value detection algorithm of slow frequency hopping signal based on plane sheet barycenter and various peak value detection algorithms.

[0005] The current method for estimating the frequency hopping period according to the time-frequency peak value mainly includes a successive CT (center time value of frequency hopping pulse) difference histogram method. This method estimates the frequency hopping period of multiple signals by using the time difference between the frequency hopping pulse peak values, but is only applicable to the case that the hopping speed is fixed and the signal detection condition is good (the number of missed detection hops is very small). If the signal detection effect is poor, the multiple frequency hopping pulses are missed, and the wrong estimation value is easily obtained. In addition, this algorithm needs to calculate multiple difference values, and the calculation amount is large. SUMMARY

[0006] The application proposes a frequency hopping period estimation method based on FFT transformation for the frequency hopping signal frequency hopping period estimation problem.

[0007] The technical scheme adopted by the application is:

[0008] The peak value detection algorithm is used to find the time-frequency peak value corresponding to each frequency hopping pulse of the frequency hopping signal, the position of the time-frequency peak value in the time-frequency matrix is recorded, the time parameter in the time-frequency peak value is extracted, the time center sequence of the frequency hopping pulse is created by using the time parameter, the FFT spectrum corresponding to the sequence is obtained by performing FFT transformation on the sequence, and the frequency hopping period parameter is estimated from the FFT spectrum diagram.

[0009] In the set time-frequency band, one frequency hopping signal and n background signals are obtained as

[0010] y n =y TH +y bs1 +y bS2 +…+y bsn

[0011] Wherein, y TH is the frequency hopping signal, and y bsn is the background signal.

[0012] The time-frequency representation of the received signal y n is performed by using a spectrum diagram, and a time-frequency matrix S n is obtained, the spectrum diagram is defined as the square of the short-time Fourier transform module, and the time-frequency spectrum diagram is represented as:

[0013]

[0014]

[0015] The time length of the received signal y n is t, the row length of the obtained time-frequency matrix S n represents the sampling number L t in time, and the time resolution parameter T r in the time-frequency spectrum diagram is:

[0016]

[0017] The time-frequency peak value matrix TF of the center time-frequency peak value of each frequency hopping pulse in the frequency hopping signal is obtained by using a peak value detection algorithm.

[0018] A time sequence CT all of which is 0 is created, the length of the time sequence CT is the same as that of the time-frequency matrix S n , the position of the time center value in the time-frequency peak value matrix TF in the time sequence CT is set to 1, and the positions of the other time center values are set to 0, so that the time sequence CT is the frequency hopping pulse time center sequence containing the time center information of the frequency hopping pulse.​

[0019] FFT transform the time center sequence CT of the frequency hopping pulse to obtain an FFT spectrum P after FFT transform CT ;

[0020] FFT transform the obtained time center sequence P CT Extract the peak value, set the peak value prominence as a threshold ε1, and consider the peak value exceeding the threshold as the peak value in the FFT spectrum P CT Extract the peak value, set the peak value prominence as a threshold ε1, and consider the peak value exceeding the threshold as the peak value in the FFT spectrum P CT ;

[0021] a CT The first peak value in the peak value abscissa sequence a

[0022] Obtain the reciprocal of the obtained pulse hopping frequency, which is the time sampling point number corresponding to the hopping period T H The value of the hopping period T

[0023]

[0024] The present application has the following beneficial effects:

[0025] Compared with the successive CT difference histogram method, the hopping period estimation method based on FFT transform can obtain accurate estimation values in the case of poor detection condition (more missed detection hops), and the method only needs to perform FFT calculation once, has small calculation amount, is simple, and has good effect. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 It is a flow chart of the hopping period estimation method based on FFT transform;

[0027] Figure 2 It is a time-frequency diagram during frequency hopping signal detection;

[0028] Figure 3 It is the FFT spectrum after FFT transform of the time center sequence of the frequency hopping pulse after frequency hopping signal processing;

[0029] Figure 4 It is a time-frequency diagram during frequency hopping and comb spectrum signal detection;

[0030] Figure 5 It is the FFT spectrum after FFT transform of the time center sequence of the frequency hopping pulse after frequency hopping and comb spectrum signal processing. DETAILED DESCRIPTION

[0031] Referring to Figure 1 , the method specifically comprises the following steps:

[0032] (1) detect 1 frequency hopping signal and n background signals in a time-frequency band

[0033] y n = y TH + y bs1 + y bs2 + … + y bsn (1)

[0034] (2) Based on the short-time Fourier transform of the spectrum effect is good, can be clearly identified signal, and in engineering has been widely used, so use the spectrum (Spectrogram, SP) will intercept signal y n Time-frequency representation, get time-frequency matrix S n , the so-called spectrum definition is the square of the short-time Fourier transform, time-frequency spectrum is expressed as:

[0035]

[0036] (3) Given the intercept signal y n The length of time t, (2) obtained in the time-frequency matrix S n The length of the row represents the number of samples L t in time, the time resolution parameter T r in the time-frequency spectrum is:

[0037]

[0038] (4) Through the peak value detection algorithm based on two-dimensional quadratic difference of fast frequency hopping signal, based on the slow frequency hopping signal time-frequency peak value detection algorithm of the center of gravity of the plane sheet, etc. Find the center of each hop time-frequency peak value of the frequency hopping pulse in the frequency hopping network station signal, get the time-frequency peak matrix TF.

[0039] (5) Create a time series CT all 0, the length of the sequence is the same as the length of the time-frequency matrix S n , the time center value in the time-frequency peak matrix TF is set to 1, and the rest is set to 0. At this time, the time series CT is the frequency hopping pulse time center sequence containing the time center information of the frequency hopping pulse.

[0040] (6) FFT transform the frequency hopping pulse time center sequence CT, get the FFT spectrum P CT after FFT transform.

[0041] (7) Extract the peak value of the time center sequence FFT spectrum P CT obtained in (6), set the peak prominence to threshold ε1, and consider it as the FFT spectrum P CTThe peak value, peak prominence, is used to measure the prominence of a peak (its prominence relative to the positions of other peaks). Based on experience, a value of 0.33 is set. The abscissas of all peaks are collected to form a peak abscissa sequence 'a'. CT .

[0042] (8)a CT The first peak on the horizontal axis corresponds to the frequency hopping pulse switching frequency a1 (the number of hopping cycles corresponding to a single sampling point). To reduce the error rate, a1 can also be adjusted. CT The difference sequence b is obtained by performing difference processing. CT b CT The modal value in the value corresponds to the pulse hopping frequency a1 of the frequency hopping network station.

[0043] (9) Take the reciprocal of the pulse transition frequency obtained in (8), which is the number of time sampling points corresponding to the jump period, T. H The value is:

[0044]

[0045] In one simulation example, a frequency-hopping network station is designed as a frequency-hopping reconnaissance target, with a frequency range of 12-16MHz, a hopping rate of 1000 hops / s, and 64 hopping frequency points.

[0046] (1) The frequency hopping signal detected in the 12-16MHz frequency band is y n .

[0047] (2) The detected signal y n Time-frequency representation is performed, and the time-spectrum diagram is as follows: Figure 2 As shown, the corresponding time-frequency matrix S n :

[0048]

[0049] (3) The received signal y is known. n The time length is t = 64 ms, and the time-frequency matrix S obtained in (2) is n The line length L represents the number of samples in time. t =1919, the time resolution parameter T in the time-space spectrum is... r That is:

[0050]

[0051] (4) Find the center time-frequency peak value of each hop of the frequency hopping pulse in the frequency hopping network signal by using the time-frequency peak detection algorithm, and obtain the time-frequency peak matrix TF.

[0052] (5) Create a time series CT consisting entirely of zeros, the length of which is equal to the time-frequency matrix S. nThe length is 1919, the time center value in the time-frequency peak matrix TF is set to 1, and the remaining positions are set to 0. At this time, the time sequence CT is the frequency hopping pulse time center sequence containing the time center information of the frequency hopping pulse.

[0053] (6) The FFT transform of the frequency hopping pulse time center sequence CT is performed to obtain the FFT spectrum P of the time center sequence CT , as shown in the following formula. Figure 3

[0054] (7) The time center sequence FFT spectrum P obtained in (6) CT extracts the peak value, and the peak value prominence threshold ε2 is set to 0.33, Figure 3 The positions of the three triangular marks are the positions of the peak values, and the abscissas of all the peak values are collected to form the peak abscissa sequence a CT :

[0055] a CT = [0.0334, 0.0667, 0.1001, 0.1334, 0.1668, 0.2001, 0.2335, 0.2668, 0.3002, 0.3335, 0.3669, 0.4002, 0.4336, 0.4669 (7)

[0056] (8) The difference sequence b CT is obtained by performing difference processing on a CT :

[0057] b CT = [0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334] (8)

[0058] The mode value in b CT corresponds to the frequency hopping network station pulse hopping frequency a1=0.0334.

[0059] (9) The reciprocal of the pulse hopping frequency obtained in (8) is the time sampling point number corresponding to the hop period, and the value of the hop period T H is:

[0060]

[0061] ​In order to verify the detection ability of the FFT transform based hopping period estimation method in the case of missing hops, in another simulation example, a frequency hopping network station is designed as a frequency hopping reconnaissance target object, with a frequency range of 12-16 MHz, a hopping speed of 1000 hops / s, and 64 frequency hopping points; 32 frequency points are randomly selected from the 64 frequency points of the frequency hopping network station to transmit comb spectrum signals as interference signals.

[0062] (1) The frequency hopping signals and 32 comb spectrum signals in the 12-16 MHz time-frequency band are detected as:

[0063] y n = y TH +y bs (10)

[0064] (2) The detected signal y n is represented in time-frequency, and the time-frequency spectrum is shown in Figure 4 , and the corresponding time-frequency matrix S n is:

[0065]

[0066] (3) The time length of the detected signal y n is t=64 ms, and the row length of the time-frequency matrix S n obtained in (2) represents the number of samples L t =1919 in time, and the time resolution parameter T r in the time-frequency spectrum in this time is:

[0067]

[0068] (4) The time-frequency peak detection algorithm is used to find the center time-frequency peak of each hop of the frequency hopping pulse in the frequency hopping network station signal, and the time-frequency peak matrix TF is obtained.

[0069] (5) A time sequence CT is created, which is all 0, and the length of the sequence is the same as the length of the time-frequency matrix S n , i.e. 1919, and the time center value in the time-frequency peak matrix TF is set to 1 in the corresponding position of the time sequence CT, and the remaining positions are set to 0. At this time, the time sequence CT is the frequency hopping pulse time center sequence containing the time center information of the frequency hopping pulse.

[0070] (6) The FFT transform is performed on the frequency hopping pulse time center sequence CT to obtain the FFT spectrum P CT after FFT transform, as shown in Figure 5 .

[0071] (7) The peak value is extracted from the time center sequence FFT spectrum P CT obtained in (6), and the peak prominence threshold ε2 is set to 0.33,Figure 5 The positions of the middle triangular marks are all the positions of the peaks, and the abscissas of all the peaks are collected to form a peak abscissa sequence:

[0072] a CT = [0.0334, 0.0667, 0.1001, 0.1334, 0.1668, 0.2001, 0.2335, 0.2668, 0.3002, 0.3335, 0.3669, 0.4002, 0.4336, 0.4669] (13)

[0073] (8) The difference sequence b CT is obtained by differentiating a CT :

[0074] b CT = [0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334, 0.0334] (14)

[0075] The mode value in b CT is 0.0334, which corresponds to the frequency a1 of the pulse hopping of the frequency hopping network station.

[0076] (9) The reciprocal of the pulse hopping frequency obtained in (8) is the time sampling point number corresponding to the hopping period, and the value of the hopping period T H is:

[0077]

[0078] Hopping period estimation effect:

[0079] Without adding the comb spectrum signal, the hopping period value estimated by the hopping period estimation method based on FFT transformation is 1000us, which is the same as the actual hopping period of the frequency hopping network station, while the hopping period value estimated by the successive CT difference histogram method under the same conditions is 1001us, which is due to the influence of the time resolution of the successive CT difference method, while the former reduces the influence of the insufficient time resolution to a certain extent; after adding the comb spectrum signal, the hopping period value estimated by the hopping period estimation method based on FFT transformation is still 1000us, while the hopping period value estimated by the successive CT difference histogram method is 23010us, which is far from the actual value. The simulation analysis results prove that the method can accurately estimate the hopping period parameter of the frequency hopping signal under the condition of poor detection effect, and confirm the effectiveness of the hopping period estimation method based on FFT transformation.

Claims

1. A method for estimating a hop period based on FFT transform, characterized by, Comprise: In the set time-frequency band, 1 frequency hopping signal and n background signals are obtained y n = y TH + y bs1 + y bs2 +... + y bsn wherein y TH is the frequency hopping signal, y bsn is the background signal; Using a spectrogram on the intercepted signal y n Performing a time-frequency representation to obtain a time-frequency matrix S n The spectrogram is defined as the square of the modulus of the short-time Fourier transform and the time-frequency spectrogram is represented as: The intercepted signal y is known n The time length is t, and a time-frequency matrix S is obtained n The row length of the time-frequency matrix S represents the number of samples L in time t The time resolution parameter T in the time-frequency spectrum is r ​ The center time-frequency peak of each hop of the frequency hopping pulse in the frequency hopping network station signal is found by the peak detection algorithm, and a time-frequency peak matrix TF is obtained; A time sequence CT of all 0 is created, the length of which is the same as that of the time-frequency matrix S n The time center value in the time-frequency peak matrix TF is set to 1 in the position corresponding to the time sequence CT, and the rest is set to 0. At this time, the time sequence CT is the frequency hopping pulse time center sequence containing the frequency hopping pulse time center information. The FFT transform is performed on the time center sequence CT of the frequency hopping pulse to obtain the FFT spectrum P after the FFT transform CT ; The time center sequence FFT spectrum P is obtained CT The peak value is extracted, and a peak prominence is set as a threshold ε1. The peak value exceeding the threshold is considered as a peak in the FFT spectrum P CT The abscissa of all the peak values is collected to form a peak abscissa sequence a CT ; a CT The first peak abscissa in the graph corresponds to the frequency a1 of the frequency hopping pulse. The reciprocal of the obtained pulse jump frequency is the time sampling point number corresponding to the jump period T H The value is:

Images