Fast implementation method and device for long-time coherent accumulation of GPS external source signal

Abstract

The application discloses a kind of GPS external radiation source signal long time coherent accumulation fast implementation method and its device, comprising: obtaining cross-correlation spectrum, obtain distance and doppler frequency migration quantity;Parabolic sampling is carried out to cross-correlation spectrum, and sampling sequence is obtained;Construct target function, carry out parabolic sampling to cross-correlation spectrum, frequency compensation and Fourier transform, obtain peak value and corresponding speed;Using improved particle swarm optimization algorithm, the target function value is maximized, and the estimated equivalent speed and the estimated equivalent acceleration are obtained, and the distance and doppler frequency migration quantity are corrected, and the corrected cross-correlation spectrum is obtained;Along the slow time Fourier transform is carried out to the corrected cross-correlation spectrum, is converted into distance and doppler frequency spectrum, and peak detection is carried out, and peak signal-to-noise ratio, distance difference and speed are obtained.The application can greatly reduce the amount of calculation and running time, improve the correctness of improved particle swarm optimization algorithm estimation migration correction parameter, and can detect multiple targets.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of signal processing, and particularly relates to a fast implementation method and device for long-time coherent accumulation of GPS external signal. BACKGROUND

[0002] The global positioning system has the advantages of global coverage and stable operation. Among them, the GPS external source radar is deployed in key areas such as coastal ports and islands for silent air and sea monitoring, but the power of the GPS satellite signal reaching the ground is too small, and the signal-to-noise ratio (SNR) of the target echo signal is extremely low, so long-time coherent accumulation is needed to make the signal energy exceed the detection threshold. For example, in the traditional pulse Doppler processing method, the one-dimensional sequence of continuous direct wave and target echo signal is first equivalent pulsed into a two-dimensional matrix according to a 1ms period, then the direct wave and target echo equivalent pulse signals are cross-correlated along the fast time dimension, and finally the range-doppler frequency (RD) spectrum is obtained by performing Fourier transform (FFT) on the cross-correlation spectrum along the slow time dimension. In this process, the coherent accumulation gain is theoretically proportional to the accumulation time, but as the accumulation time increases, the range and Doppler frequency migration caused by target motion intensifies, resulting in dispersion of the RD spectrum peak energy, serious loss of accumulation gain, and even the gain decreases after the accumulation time exceeds a certain limit.

[0003] In order to make the accumulated post-peak energy concentrated and reach the maximum, migration correction is essential, and the existing methods can be divided into parameterized methods and non-parameterized methods. The commonly used non-parameterized methods are Keystone Transform (KT) and Radon-Fourier Transform (RFT), which can realize first-order distance migration blind correction under the condition that the target prior motion information is unknown, but the second-order distance curvature and Doppler frequency migration are not corrected. The parameterized methods include Wigner-Ville Distribution (WVD), Lv's Distribution (LVD), Fractional Fourier Transform (FrFT) and the like, which regard the cross-correlation spectrum slow time sequence as a Linear Frequency Modulation (LFM) signal, estimate the center frequency and Doppler frequency modulation of two parameters, and realize accurate distance and Doppler frequency migration correction. Since the peak energy maximization belongs to the optimization problem, the group intelligence optimization algorithm can be used to estimate the parameters, but the existing Particle Swarm Optimization (PSO) algorithm can estimate the migration correction parameters, but there are problems of large amount of calculation, long running time, possible convergence to local optimum and inability to search for weak target parameters in multi-target detection scenarios.

[0004] Therefore, it is urgent to provide a fast implementation method of GPS external radiated source signal long time coherent accumulation to improve the above defects. SUMMARY

[0005] In order to solve the above problems existing in the prior art, the application provides a fast implementation method of GPS external radiated source signal long time coherent accumulation and a device thereof. The technical problems to be solved by the application are solved by the following technical solutions:

[0006] In the first aspect, the application provides a fast implementation method of GPS external radiated source signal long time coherent accumulation, comprising:

[0007] Obtaining a cross-correlation spectrum, converting a distance difference related to the cross-correlation spectrum, so that the distance difference is related to an initial distance difference, an equivalent velocity and an equivalent acceleration, and obtaining a distance and Doppler frequency migration;

[0008] Constructing a target function, and inputting the initial distance difference, the equivalent velocity and the equivalent acceleration, performing parabolic sampling and frequency compensation on the cross-correlation spectrum to obtain a sampling sequence; performing Fourier transform on the sampling sequence to output a peak value of the cross-correlation spectrum and a velocity corresponding to the peak value;

[0009] An improved particle swarm optimization algorithm is used to maximize the objective function value, so as to obtain the estimated equivalent velocity and the estimated equivalent acceleration;

[0010] According to the estimated equivalent velocity and the estimated equivalent acceleration, the distance and Doppler frequency migration quantity is corrected to obtain a corrected cross-correlation spectrum; the corrected cross-correlation spectrum is subjected to Fourier transform along the slow time to be converted into a distance and Doppler frequency spectrum; and the distance and Doppler frequency spectrum is subjected to peak value detection to obtain a peak value signal-to-noise ratio, a distance difference and a velocity.

[0011] In a second aspect, the application further provides a device for fast implementation of long-time coherent accumulation of GPS external source signals, comprising:

[0012] A data acquisition unit is configured to acquire a cross-correlation spectrum, convert a distance difference related to the cross-correlation spectrum so that the distance difference is related to an initial distance difference, an equivalent velocity and an equivalent acceleration, and acquire a distance and Doppler frequency migration quantity.

[0013] A data processing unit is configured to construct an objective function, input the initial distance difference, the equivalent velocity and the equivalent acceleration, perform parabolic sampling and frequency compensation on the cross-correlation spectrum to obtain a sampling sequence, and perform Fourier transform on the sampling sequence to output a peak value of the cross-correlation spectrum and a velocity corresponding to the peak value.

[0014] A data optimization unit is configured to use an improved particle swarm optimization algorithm to maximize the objective function value, so as to obtain the estimated equivalent velocity and the estimated equivalent acceleration.

[0015] A result acquisition unit is configured to correct the distance and Doppler frequency migration quantity according to the estimated equivalent velocity and the estimated equivalent acceleration to obtain a corrected cross-correlation spectrum, perform Fourier transform on the corrected cross-correlation spectrum along the slow time to be converted into a distance and Doppler frequency spectrum, and perform peak value detection on the distance and Doppler frequency spectrum to obtain a peak value signal-to-noise ratio, a distance difference and a velocity.

[0016] The application has the following beneficial effects:

[0017] The GPS external source signal long-time coherent accumulation fast implementation method and device provided by the application can realize accurate migration correction of distance and Doppler frequency by performing parabolic sampling and frequency compensation on the cross-correlation spectrum in the objective function, and then performing Fourier transform on the one-dimensional sampling sequence, so that the calculation amount and running time are greatly reduced; the adaptive weight coefficient and mutation mechanism are introduced into the improved particle swarm optimization algorithm process, so that the improved particle swarm optimization algorithm is prevented from falling into local optimum, and the correctness of the improved particle swarm optimization algorithm in estimating the migration correction parameters is improved; the forbidden zone is set in the objective function, and the forbidden zone is constantly added according to the searched parameter results in the process of repeatedly running the improved particle swarm optimization algorithm, so that multi-target detection is realized.

[0018] The application will be described in further detail below with reference to the drawings and embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 is a flow chart of the fast implementation method of long-time coherent accumulation of GPS external source signal provided by the embodiments of the application;

[0020] Figure 2 is a schematic diagram of the fast implementation method of long-time coherent accumulation of GPS external source signal provided by the embodiments of the application;

[0021] Figure 3 is a schematic diagram of constructing the target function provided by the embodiments of the application;

[0022] Figure 4 is a schematic diagram of the improved particle swarm optimization algorithm provided by the embodiments of the application;

[0023] Figure 5 a~5f is a schematic diagram of processing the cross-correlation spectrum result using the pulse Doppler and PSO+FFT method provided by the embodiments of the application;

[0024] Figure 6 a~6f is a schematic diagram of the RD spectrum result after PSO+FFT in the multi-target scenario provided by the embodiments of the application. DETAILED DESCRIPTION

[0025] The application will be described in further detail below with reference to the drawings and embodiments.

[0026] The existing PSO migration correction parameter estimation algorithm needs to reconstruct the entire cross-correlation spectrum two-dimensional matrix according to the two motion parameters of velocity and acceleration, and then perform Fourier transform along the slow time dimension to output the distance and Doppler frequency spectrum peak energy. The number of calls of the target function is the product of the iteration number and the particle number, so the existing particle swarm optimization algorithm has the problems of large amount of calculation, long running time, and being prone to falling into local optimum. Because the optimization target of the PSO algorithm is to maximize the RD spectrum peak energy, in the GPS external source radar multi-target detection scenario, the PSO algorithm can only search for the motion parameters corresponding to the target with the strongest energy, and ignores other targets with weaker energy.

[0027] Therefore, the application provides a GPS external radiation source signal long-time coherent accumulation fast implementation method, which has the following three advantages.

[0028] Please refer to Figure 1 , Figure 1 is a flow chart of the GPS external radiation source signal long-time coherent accumulation fast implementation method provided by the application, Figure 2 is a schematic diagram of the GPS external radiation source signal long-time coherent accumulation fast implementation method provided by the application, and the GPS external radiation source signal long-time coherent accumulation fast implementation method provided by the application comprises the following steps.

[0029] S101, the cross-correlation spectrum is obtained, the distance difference related to the cross-correlation spectrum is converted, the distance difference is related to the initial distance difference, the equivalent velocity and the equivalent acceleration, and the distance and Doppler frequency migration amount are obtained.

[0030] Specifically, in the embodiment, the continuous wave signal emitted by the GPS external radiation source radar is represented as:

[0031] S t (t)=A·C(t)·D(t)·exp(j2πf L t);

[0032] Wherein, A represents the signal amplitude, C(t) represents C / A code, which belongs to a pseudo-random spread spectrum sequence, the code rate is 1.023MHz, and the period is 1ms; D(t) represents BPSK modulated navigation text data, and the carrier frequency f L is 1575.42MHz.

[0033] According to the continuous wave signal, the continuous direct wave signal and the target echo signal are obtained, and the one-dimensional sequence of the continuous direct wave signal and the target echo signal is equivalent pulsed into a two-dimensional matrix according to a preset period, and the expressions are as follows:

[0034]

[0035] Wherein, S ref (τ,η) represents the two-dimensional matrix corresponding to the continuous direct wave signal, S r(τ,η) represents a two-dimensional matrix corresponding to the target echo signal, τ∈[0,T] represents fast time, T represents a preset period, which can be 1 ms, η∈[0,T D ] represents slow time, and η takes an integer multiple of T, T D represents total accumulation time, A ref represents the amplitude of the continuous direct wave signal, A r represents the amplitude of the target echo signal, u(·) represents the continuous wave signal envelope transmitted by the GPS external radiation source radar, that is, C(·)D(·), L(η) represents the distance between the receiver of the GPS external radiation source radar and the satellite, R t (η) represents the distance between the satellite and the target, R r (η) represents the distance between the receiver of the GPS external radiation source radar and the target, and c represents the speed of light.

[0036] Along the fast time dimension, the two-dimensional matrix corresponding to the continuous direct wave signal and the two-dimensional matrix corresponding to the target echo signal are multiplied in the frequency domain to realize cross-correlation operation, and a cross-correlation spectrum is obtained, and the expression is as follows:

[0037]

[0038] ΔR(η)=R t (η)+R r (η)-L(η);

[0039] Wherein, represents convolution, U ref (f τ ,η) represents a direct wave frequency domain signal, U r * (f τ ,η) represents the conjugate of the target echo frequency domain signal, and the product is converted into a time domain cross-correlation spectrum by inverse Fourier transform;

[0040]

[0041] Wherein, PC(τ,η) represents the cross-correlation spectrum, CF(·) represents the periodic correlation function of the C / A code of the GPS external radiation source radar, and ΔR(η) represents the distance difference varying with slow time.

[0042] In this embodiment, the distance difference related to the cross-correlation spectrum is converted, so that the distance difference is related to the initial distance difference, the equivalent velocity and the equivalent acceleration, and the distance and the Doppler frequency migration are obtained, including:

[0043] The distance difference varying with slow time is Taylor expanded at η=0, and converted into a quadratic polynomial, and the expression is as follows:

[0044]

[0045] wherein, R0 represents an initial distance difference, v0 represents an equivalent velocity, and a0 represents an equivalent acceleration;

[0046] According to the quadratic polynomial, the distance and the Doppler frequency migration quantity are obtained, and the expression is:

[0047]

[0048] wherein, RM(η) represents the distance migration quantity, and DFM(η) represents the Doppler frequency migration quantity.

[0049] S102, a target function is constructed, and an initial distance difference, an equivalent velocity and an equivalent acceleration are input, parabolic sampling and frequency compensation are performed on the cross-correlation spectrum, and a sampling sequence is obtained; Fourier transform is performed on the sampling sequence, and a peak value of the cross-correlation spectrum and a velocity corresponding to the peak value are output.

[0050] Specifically, please refer to Figure 3 , Figure 3 is a schematic diagram of constructing a target function provided by the embodiment of the application, in the embodiment, the cross-correlation spectrum is subjected to parabolic sampling, each sampling point is multiplied by to obtain a sampling sequence s(η), and the expression is:

[0051]

[0052] wherein, represents an estimated initial distance difference, represents an estimated velocity, represents an estimated acceleration, c represents the speed of light, λ represents the wavelength, and η represents the slow time;

[0053] Fourier transform is performed on the sampling sequence s(η) and is taken as a module to obtain a spectrum amplitude |U(f d )|, f d represents the Doppler frequency, the Doppler frequency f d is fixed in the range of -500-500Hz, and the Doppler frequency unit resolution is 1 / T D .

[0054] The maximum value of the spectrum amplitude is taken, and the peak position of the cross-correlation spectrum is obtained; the frequency corresponding to the peak position of the cross-correlation spectrum is converted into a velocity v d =f d λ, and simultaneously, the peak value E of the cross-correlation spectrum is obtained to construct a target function f(X), and the parameters of the target function f(X) are The expression is:

[0055]

[0056] where FFT(·) represents a Fourier transform.

[0057] In this embodiment, the cross-correlation spectrum is parabolic sampled, including:

[0058] For the fast time unit of the cross-correlation spectrum, the data of the mth fast time unit is taken as the sampling data;

[0059] where the data of the mth fast time unit is rounded, F S represents a sampling rate, and the fast time unit resolution of the cross-correlation spectrum is 1 / F S .

[0060] For the slow time unit of the cross-correlation spectrum, η=nT is taken as the sampling data; where n represents a pulse sequence number, T represents a preset period, and the sampling sequence length is the pulse number N p =T D / T.

[0061] S103, using the improved particle swarm optimization algorithm, maximizing the objective function value, obtaining the estimated equivalent velocity and the estimated equivalent acceleration.

[0062] Specifically, please refer to Figure 4 , Figure 4 is a schematic diagram of the improved particle swarm optimization algorithm provided by the embodiment of the application, in this embodiment, first, set the basic parameters, the total number of particles n=30, the maximum iteration number N=150, and the upper and lower bounds of the distance difference-speed-acceleration three-dimensional search space are 0.5-5km, -25-25m / s and -1-1m / s 2 .

[0063] Randomly initialize the position and the velocity of n particles within the preset limit, i=1, 2, …, n represents the particle sequence number, and k represents the iteration number, the historical optimal position of each particle is initialized as

[0064] Calculate the objective function of the ith particle determine the particle swarm historical global optimal position and the objective function of the particle swarm where the expression of the objective function of the particle swarm is:

[0065]

[0066] Calculate the distance of the position of each particle to the particle swarm historical global optimal position in the kth iteration process And according to the way from far to near, the particle is sorted, and the distance order ranking is obtained;

[0067] Determine Whether to converge, where the convergence condition is:

[0068]

[0069] If not convergent, continue to update the position of the particle Speed And the historical optimal position of the particle And determine whether the iteration number k+1 reaches the maximum iteration number N; if yes, calculate the particle swarm historical global optimal position And the objective function of the particle swarm And output the maximum swarm historical global optimal position X gbest And the objective function f(X gbest ) of the particle swarm; if not, continue to update the position of the particle Speed And the historical optimal position of the particle

[0070] If convergent, record the particle swarm historical global optimal position And the objective function of the particle swarm And select any particle to carry out mutation operation; after mutation operation, continue to update the position of the particle Speed And the historical optimal position of the particle And determine whether the iteration number k+1 reaches the maximum iteration number N; if yes, calculate the particle swarm historical global optimal position And the objective function of the particle swarm And output the maximum swarm historical global optimal position X gbest And the objective function f(X gbest ) of the particle swarm; if not, continue to update the position of the particle Speed And the historical optimal position of the particle

[0071] Further, update the position of the particle Speed And the historical optimal position of the particle Including:

[0072] Update the speed of the particle, and the expression is:

[0073]

[0074] wherein c1 and c2 represent acceleration coefficients, and r1, r2 ∈ [0, 1] represent random numbers, denotes an adaptive weight coefficient, and its expression is:

[0075]

[0076] wherein, denotes the ranking of particle i in the distance order, and the distance is closer, then is smaller, and ω min and ω max are respectively 0.4 and 0.9;

[0077]

[0078] wherein v max denotes the upper limit of particle velocity, and if it is out of bounds, then is set to -v max or v max , and v max is set to 0.1 times the difference between the upper and lower limits;

[0079] The position of the particle is updated as:

[0080]

[0081] The historical optimal position of the particle is updated as:

[0082]

[0083] Further, any particle is selected for mutation operation, including:

[0084] The selection strategy includes: in the early stage of iteration, particles with a distance order ranking less than k·n / N are selected for mutation operation; and in the later stage of iteration, particles with a distance order ranking greater than k·n / N are selected for mutation operation.

[0085] The mutation operation includes: the position and the historical optimal position of the selected particle are updated to a specific value, and the velocity is randomly reset, and the specific value is selected The estimated equivalent velocity is updated to the objective function of the particle swarm The output v d is added by a random disturbance ε, that is,

[0086] S104, according to the estimated equivalent velocity and the estimated equivalent acceleration, correcting the range and Doppler frequency migration, obtaining a corrected cross-correlation spectrum; performing Fourier transform on the corrected cross-correlation spectrum along the slow time, and converting into a range and Doppler frequency spectrum; performing peak value detection on the range and Doppler frequency spectrum, and obtaining a peak signal-to-noise ratio, a range difference and a velocity.

[0087] Specifically, in the embodiment, for the cross-correlation spectrum, each equivalent pulse is cyclically shifted left by m units along the fast time dimension and multiplied by to correct the range and Doppler frequency migration, and the expression is:

[0088]

[0089] m is rounded.

[0090] Then, Fourier transform (FFT) is performed along the slow time dimension, and the RD spectrum is converted. Finally, two-dimensional constant false alarm rate (2D-CFAR) detection is performed on the RD spectrum, and a peak SNR, a range and a velocity are output.

[0091] To sum up, the application provides a fast implementation method of long-time coherent accumulation of GPS external radiated source signals, and compared with the existing particle swarm algorithm process, an adaptive weight coefficient ω i is used. gbest The closer the particle position is to the historical optimal position X i of the group, the smaller ω τ is, so that the global search ability of the particle at a long distance is strong, and the convergence speed of the particle at a short distance is fast. In order to avoid premature convergence to a local optimal point, a mutation mechanism is introduced, and a particle at a short distance is selected before iteration, and its position is updated to which can jump out of the local optimum.

[0092] By analyzing and comparing the target function calculation complexity of the improved PSO algorithm and the traditional PSO algorithm, it is found that the traditional PSO algorithm only estimates two motion parameters The migration correction is realized by cyclically shifting and frequency compensating the equivalent pulses, which is equivalent to reconstructing the entire cross-correlation spectrum two-dimensional matrix, and then performing FFT conversion into the RD spectrum along the slow time dimension. τ If the fast time point number N S = F p T, the pulse number N D =T / T, the calculation complexity of the traditional PSO algorithm for migration correction is O(N τ N p ), the FFT calculation complexity is O(N τ N p logN p ), and the total calculation complexity is O(N τ Np logN p ). The improved fast PSO algorithm estimates three motion parameters The motion correction is realized by parabolic sampling and frequency compensation on the cross-correlation spectrum, and FFT is only performed on the one-dimensional sampling sequence. The calculation complexity of the improved PSO algorithm for motion correction is O(N p ), the FFT calculation complexity is O(N p logN p ), and the total calculation complexity is O(N p logN p ), so theoretically, the improved PSO algorithm is N τ times faster than the traditional PSO algorithm.

[0093] Based on the same inventive concept, the application also provides a device for fast realization of long-time coherent accumulation of GPS external source signals, which is used to realize the method for fast realization of long-time coherent accumulation of GPS external source signals provided in the above-mentioned embodiments of the application. The embodiments of the method are described above and will not be repeated here. The device comprises:

[0094] a data acquisition unit configured to acquire a cross-correlation spectrum, convert a distance difference related to the cross-correlation spectrum so that the distance difference is related to an initial distance difference, an equivalent velocity and an equivalent acceleration, and acquire a distance and a Doppler frequency migration amount;

[0095] a data processing unit configured to construct a target function, input the initial distance difference, the equivalent velocity and the equivalent acceleration, perform parabolic sampling and frequency compensation on the cross-correlation spectrum to obtain a sampling sequence, and perform Fourier transform on the sampling sequence to output a peak value of the cross-correlation spectrum and a velocity corresponding to the peak value;

[0096] a data optimization unit configured to use an improved particle swarm optimization algorithm to maximize the value of the target function and obtain an estimated equivalent velocity and an estimated equivalent acceleration;

[0097] a result acquisition unit configured to correct the distance and the Doppler frequency migration amount according to the estimated equivalent velocity and the estimated equivalent acceleration to obtain a corrected cross-correlation spectrum, perform Fourier transform on the corrected cross-correlation spectrum along a slow time to convert the corrected cross-correlation spectrum into a distance and Doppler frequency spectrum, and perform peak value detection on the distance and Doppler frequency spectrum to obtain a peak signal-to-noise ratio, a distance difference and a velocity.

[0098] In an alternative embodiment of the application, the effect of the method for fast realization of long-time coherent accumulation of GPS external source signals provided in the above-mentioned embodiments is verified through a simulation experiment, specifically as follows:

[0099] Simulation Experiment One

[0100] The scenario of simulation experiment one is single target detection, the initial range difference of target is 2000.1m, the equivalent velocity is-19.07m / s, the negative value indicates that the target is approaching the radar receiver, the equivalent acceleration is 0.045m / s 2 , the target echo signal SNR is set to-40dB, the sampling rate is 20.46MHz, and the total accumulation time is 10s.

[0101] The cross-correlation spectrum is processed by using the pulse Doppler and PSO+FFT method respectively, and the results are shown in Figure 5 a~5f and Table 1, the search parameters of the PSO algorithm are [2009.0m, -17.665m / s, 0.04685m / s 2 ] respectively. 5a is the original cross-correlation spectrum, 5b is the cross-correlation spectrum after the migration correction according to the estimated parameters of the PSO, which shows that the estimated parameters of the PSO are basically correct, and the migration correction makes the peaks of the cross-correlation spectrum basically aligned, 5c is the RD spectrum after the pulse Doppler processing, 5d is the RD spectrum after the PSO+FFT processing, which shows that the migration correction makes the peak energy of the RD spectrum concentrated, 5e is the Doppler frequency dimension profile of the RD spectrum after the pulse Doppler processing, and 5f is the Doppler frequency dimension profile of the RD spectrum after the PSO+FFT processing, which shows that the PSO+FFT can obtain a higher peak SNR, and Table 1 shows that the SNR improvement value is 6.5dB.

[0102] Table 1: RD spectrum detection results obtained by pulse Doppler and PSO+FFT

[0103] Peak SNR (dB) Distance (m) Velocity (m / s) Pulse Doppler 15.5 1964.8 -18.471 PSO + FFT 22.0 2008.8 -18.547

[0104] The traditional PSO and the improved PSO algorithm are used for migration correction parameter estimation respectively, and the running time of the program is tested as shown in Table 2. The speed of the improved PSO algorithm is nearly 160 times faster than that of the traditional PSO algorithm.

[0105] Table 2: Program test results of migration correction parameter estimation by traditional PSO and improved PSO algorithm

[0106]

[0107]

[0108] The total number of particles, the maximum number of iterations, the upper and lower bounds of the search space, the acceleration coefficient and the maximum speed limit of the traditional PSO and the improved PSO algorithm are kept consistent, the fixed weight coefficient 0.6 is used for the traditional PSO algorithm, the adaptive weight coefficient and the mutation mechanism are introduced for the improved PSO algorithm, and the Monte Carlo experiment of migration correction parameter estimation is carried out, which is repeated for 500 times respectively. If the final result f(X gbest) is regarded as error, the statistical accuracy rate is shown in Table 3, which shows that the adaptive weight coefficient and variation mechanism proposed in the application can effectively avoid the PSO algorithm from falling into local optimum, and improve the accuracy rate by 3.4%.

[0109] Table 3 Accuracy rate statistics of traditional PSO and improved PSO algorithm for motion correction parameter estimation

[0110] Correct times Incorrect times Accuracy (%) Error rate (%) Traditional PSO 474 26 94.8 5.2 Improved PSO 491 9 98.2 1.8

[0111] Simulation experiment two

[0112] The scenario of simulation experiment two is multi-target detection, the motion parameters of three targets are shown in Table 4, the closer the distance, the stronger the target echo signal energy, the target echo signal SNR is set to-40dB, the sampling rate is 20.46MHz, and the total accumulation time is 10s.

[0113] Table 4 Motion parameters of three targets

[0114] Initial distance difference (m) Equivalent velocity (m / s) Equivalent acceleration (m / s 2 )]]> Target 1 1375.1 -8.57 0.075 Target 2 1541.7 -22.61 0.119 Target 3 2000.1 -19.07 0.045

[0115] The motion correction parameters searched by the PSO algorithm for the first time are [1385.4m, -8.323m / s, 0.07371m / s 2 ], the RD spectrum result of PSO+FFT is shown in Figure 6 , 6a is an overview, 6b is a local magnification at the position of target 1, and repeated running of PSO can only obtain the motion parameters of the target 1 with the strongest energy, and the energies of the other two targets on the RD spectrum are still dispersed. Therefore, a forbidden zone is set in the objective function, if the input motion parameters are in the neighborhood range of the motion parameters of the strong target searched, 0 is directly outputted, so that the same area is not searched by the PSO before and after.

[0116] If the first PSO search parameter result is [1385.4m, -8.323m / s, 0.07371m / s 2 ], the forbidden zone is set to [1385±300m, 0.0737±0.04m / s 2 ] when the second PSO search is performed, and the motion parameters [1550.6m, -21.52m / s, 0.1396m / s 2 ] of target 2 can be searched, the new RD spectrum is obtained after motion correction by FFT, and 6c is an overview of the RD spectrum, and 6d is a local magnification at the position of target 2.

[0117] By analogy, the forbidden zone is set to [1385±300m, 0.0737±0.04m / s 2 ] and [1550±300m, 0.1396±0.04m / s 2] the logic OR, the motion parameters of target 3 [2008.4m, -17.622m / s, 0.0481m / s can be searched 2 ] 6e is a total RD spectrum, and 6f is a local amplification at the position of target 3. The multi-target RD spectrum detection result after PSO+FFT is shown in Table 5.

[0118] Table 5 Multi-target RD spectrum detection result after PSO+FFT

[0119] Peak SNR (dB) Distance (m) Velocity (m / s) Target 1 20.6 1378.3 -8.569 Target 2 19.0 1554.3 -22.661 Target 3 17.2 2008.8 -19.081

[0120] In summary, several targets need to be repeatedly run several times of PSO, and the forbidden zone range is constantly added until all targets are detected.

[0121] In summary, the present application has the following beneficial effects:

[0122] The improved PSO algorithm greatly reduces the calculation amount and running time of the target function for migration correction, and the examples prove that the running time is reduced from about 50min to within 20s, and the running speed is increased by about 160 times.

[0123] The present application introduces an adaptive weight coefficient and a mutation mechanism in the PSO algorithm process, which can effectively avoid the PSO algorithm from falling into local optimum, and the Monte Carlo experiment proves that the convergence accuracy of the improved PSO algorithm is increased by 3.4%.

[0124] The present application sets a forbidden zone in the target function, which can realize multi-target detection, and the examples prove that the motion parameters of three targets can be searched by repeatedly running the improved PSO algorithm for 3 times and updating the forbidden zone for 2 times, and the peak energy is concentrated at the corresponding position on the RD spectrum.

[0125] It is to be understood that the terminology "first", "second", and the like used herein is merely intended to differentiate one element from another element, and does not imply or suggest any actual relationship or sequence between the elements. Also, the terms "comprise", "comprising", or any other variant thereof are intended to cover non-exclusive inclusions, such that an item or apparatus that comprises a list of elements does not exclude other elements not expressly listed. An element defined by the phrase "comprising a... " does not exclude the presence of additional identical elements in the item or apparatus comprising the element. The terms "connected", "coupled", and the like are not limited to direct connections or physical connections, but can include indirect connections or indirect physical connections between devices or elements such as electrical connections, whether direct or indirect. The terms "upper", "lower", "left", "right", and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are used only for the purpose of facilitating the description of the present application and simplifying the description, and thus cannot be construed as indicating or implying that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus cannot be construed as limiting the present application.

[0126] In the description of the present application, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" and the like means that the specific feature or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. The illustrative description of the above terms in the present specification does not necessarily refer to the same embodiment or example. Also, the specific feature or characteristic described can be combined in any suitable manner in one or more embodiments or examples. In addition, those skilled in the art can combine and combine different embodiments or examples described in the present specification.

[0127] The above is a further detailed description of the present application in combination with specific preferred embodiments, and cannot be considered as limiting the specific implementation of the present application to these descriptions. For those skilled in the art, without departing from the concept of the present application, a number of simple deductions or substitutions can be made, which should be considered as falling within the protection scope of the present application.

Claims

1. A fast implementation method of long-time coherent accumulation of GPS external source signal, characterized in that, The method comprises the following steps: obtaining a cross-correlation spectrum, converting a range difference related to the cross-correlation spectrum so that the range difference is related to an initial range difference, an equivalent velocity and an equivalent acceleration, and obtaining a range and Doppler frequency migration; constructing an objective function and inputting the initial range difference, the equivalent velocity and the equivalent acceleration, performing parabolic sampling and frequency compensation on the cross-correlation spectrum to obtain a sampling sequence, performing Fourier transform on the sampling sequence, and outputting a peak value of the cross-correlation spectrum and a velocity corresponding to the peak value; Using the improved particle swarm optimization algorithm, the target function value is maximized to obtain the estimated equivalent velocity and the estimated equivalent acceleration; including: randomly initializing the position and velocity of each particle within a preset limit , ;​​​​ calculating the objective function of the th particle , determining the global optimum position of the particle swarm history and the objective function of the particle swarm ; The calculation of the first iteration process is as follows The distance of each particle position to the historical global optimal position of the particle swarm And the particles are sorted in the order from far to near, and the distance order ranking is obtained. determining whether the convergence is achieved, if not, continuing to update the position of the particle velocity , and the historical optimal position of the particle , and determining whether the iteration number reaches the maximum iteration number ; if the convergence is achieved, recording the historical global optimal position of the particle swarm and the objective function of the particle swarm , and selecting any particle to perform mutation operation; wherein the position of the particle velocity , and the historical optimal position of the particle are updated updating the velocity of the particle, and the expression is: ； wherein and denotes an acceleration factor, denotes a random number, denotes an adaptive weight factor, which is expressed as ； wherein representing particles Ranking in distance order; ； wherein, represents an upper limit of the particle velocity, and if the upper limit is exceeded, the is set to or ; updating the position of the particle: ； updating the historical optimal position of the particle: ； selecting any particle to perform a mutation operation, comprising: The selection strategy includes: in the early stage of iteration, selecting particles with distance order ranking less than to perform mutation operation; in the late stage of iteration, selecting particles with distance order ranking greater than to perform mutation operation; The mutation operation includes updating the position and the historical optimal position of the selected particle to a certain value, randomly resetting the velocity, and selecting the certain value , updating the estimated equivalent velocity of the particle swarm to the objective function of the particle swarm, and outputting , adding a random disturbance to the whole , that is ; correcting the range and Doppler frequency migration according to the estimated equivalent velocity and the estimated equivalent acceleration to obtain a corrected cross-correlation spectrum, performing Fourier transform on the corrected cross-correlation spectrum along the slow time to convert it into a range and Doppler frequency spectrum, and performing peak value detection on the range and Doppler frequency spectrum to obtain a peak signal-to-noise ratio, a range difference and a velocity.

2. The method for fast implementation of long-time coherent accumulation of GPS signal from an external source according to claim 1, characterized in that, The method comprises the following steps: obtaining a cross-correlation spectrum, comprising: ； ； wherein, represents a two-dimensional matrix corresponding to the continuous direct wave signal, represents a two-dimensional matrix corresponding to the target echo signal, represents a fast time, represents a preset period, represents a slow time, and the value of is an integer multiple of , represents a total accumulation time, represents an amplitude of the continuous direct wave signal, represents an amplitude of the target echo signal, represents an envelope of the continuous wave signal emitted by the GPS external radiated source radar, represents a distance between the receiver of the GPS external radiated source radar and the satellite, represents a distance between the satellite and the target, represents a distance between the receiver of the GPS external radiated source radar and the target, represents a speed of light, represents a wavelength; obtaining a continuous direct wave signal and a target echo signal, and equivalently pulsing one-dimensional sequences of the continuous direct wave signal and the target echo signal into two-dimensional matrices according to a preset period, and the expressions are respectively: ； ； wherein, denotes a convolution, denotes a direct path frequency domain signal, denotes a conjugate of the target echo frequency domain signal, the product of which is inverse Fourier transformed into a time domain cross-correlation spectrum, denotes a range difference as a function of slow time; ； wherein, represents the cross-correlation spectrum, represents the periodic correlation function of the C / A code of the GPS SSR.

3. The method for fast implementation of long-time coherent accumulation of GPS signal from an external source according to claim 2, characterized in that, performing frequency domain conjugate multiplication on the two-dimensional matrix corresponding to the continuous direct wave signal and the two-dimensional matrix corresponding to the target echo signal along the fast time dimension to realize cross-correlation operation, and obtaining a cross-correlation spectrum, and the expression is: The distance difference as a function of slow time is Taylor expanded at place and converted into a quadratic polynomial whose expression is: ； wherein, represents the initial distance difference, represents the equivalent velocity, represents the equivalent acceleration; converting a range difference related to the cross-correlation spectrum so that the range difference is related to an initial range difference, an equivalent velocity and an equivalent acceleration, and obtaining a range and Doppler frequency migration, comprising: ； ； wherein, denotes the distance migration, denotes the Doppler frequency migration.

4. The fast implementation method of GPS signal long-time coherent accumulation of claim 1, wherein, obtaining a range and Doppler frequency migration according to a quadratic polynomial, and the expression is: constructing an objective function and inputting the initial range difference, the equivalent velocity and the equivalent acceleration, performing parabolic sampling and frequency compensation on the cross-correlation spectrum to obtain a sampling sequence; The cross-correlation spectrum is parabolic sampled, each sampling point is multiplied by , and a sampling sequence is obtained, the expression of which is ； wherein, denotes an estimated initial distance difference, denotes an estimated velocity, denotes an estimated acceleration, denotes the speed of light, denotes a wavelength, denotes a slow time; performing a Fourier transform and taking the modulus on the sampling sequence to obtain the spectral amplitudes , denotes the Doppler frequency; Taking a maximum value of the spectrum amplitude, a peak position of the cross-correlation spectrum is obtained; a frequency corresponding to the peak position of the cross-correlation spectrum is converted into a velocity Meanwhile, a peak value of the cross-correlation spectrum is obtained to construct an objective function , the objective function The parameter is and the expression is: ； wherein denotes the Fourier transform.

5. The fast implementation method of GPS signal long-time coherent accumulation of the GPS external radiation source signal according to claim 4, characterized in that, performing Fourier transform on the sampling sequence, and outputting a peak value of the cross-correlation spectrum and a velocity corresponding to the peak value, comprising: For the fast time unit of the cross-correlation spectrum, the data of the first fast time unit is taken as the sampling data; Among them, the The data for each fast time unit is Rounding down, Indicates the sampling rate; For the slow time unit of the cross-correlation spectrum, take as the sampling data; wherein, represents the pulse number, represents the preset period.

6. The method for fast implementation of GPS signal long-time coherent accumulation according to claim 1, characterized in that, performing parabolic sampling on the cross-correlation spectrum, comprising: Objective function of the swarm of particles The expression is: ； using the improved particle swarm optimization algorithm to maximize the objective function value to obtain an estimated equivalent velocity and an estimated equivalent acceleration, and further comprising: ； If the iteration number reaches the maximum iteration number , the first iteration process, the particle swarm historical global optimal position and the objective function of the particle swarm are calculated, and the maximum swarm historical global optimal position and the objective function of the particle swarm are output; if the iteration number is not the maximum iteration number , the position speed and the historical optimal position of the particle are continuously updated; After selecting any particle to perform a mutation operation, continue updating the particle's position. speed And the historical best position of the particle. And determine the number of iterations. Has the maximum number of iterations been reached? If so, then calculate the first... In the next iteration, the historical global optimal position of the particle swarm. Objective function of particle swarm It outputs the largest historical global optimal position of the group. Objective function of particle swarm If not, continue updating the particle's position. speed And the historical best position of the particle. .

7. The method for fast implementation of long-time coherent accumulation of GPS signal from an external source according to claim 1, characterized in that, the convergence condition is: For the cross-correlation spectrum, each equivalent pulse is circularly shifted left along the fast time dimension The unit is multiplied by to correct the range and Doppler frequency migration, whose expression is: ； For rounded.

8. A device for fast implementation of long-time coherent accumulation of GPS external source signal, characterized in that, correcting the range and Doppler frequency migration according to the estimated equivalent velocity and the estimated equivalent acceleration to obtain a corrected cross-correlation spectrum, comprising: comprising: a data acquisition unit configured to obtain a cross-correlation spectrum, convert a range difference related to the cross-correlation spectrum so that the range difference is related to an initial range difference, an equivalent velocity and an equivalent acceleration, and obtain a range and Doppler frequency migration; a data processing unit configured to construct an objective function and input the initial range difference, the equivalent velocity and the equivalent acceleration, perform parabolic sampling and frequency compensation on the cross-correlation spectrum to obtain a sampling sequence, perform Fourier transform on the sampling sequence, and output a peak value of the cross-correlation spectrum and a velocity corresponding to the peak value. a data optimization unit configured to maximize a target function value using an improved particle swarm optimization algorithm to obtain an estimated equivalent velocity and an estimated equivalent acceleration; a result obtaining unit configured to correct the range-Doppler migration of the estimated equivalent velocity and the estimated equivalent acceleration to obtain a corrected cross-correlation spectrum, perform Fourier transform on the corrected cross-correlation spectrum along the slow time to convert the corrected cross-correlation spectrum into a range-Doppler spectrum, and perform peak detection on the range-Doppler spectrum to obtain a peak signal-to-noise ratio, a range difference, and a velocity.

Images