A method for estimating the direction of arrival of lightning signals based on energy distribution
By determining the critical frequency point in the estimation of the lightning signal wave reach direction and segmenting the new subband based on the energy distribution, the problems of large calculation amount and low accuracy in the prior art are solved, and efficient and accurate positioning of the lightning signal is achieved.
Patent Information
- Application Number
- CN202210659040.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-06-10
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2042-06-10
AI Technical Summary
The existing lightning signal wave reach direction estimation methods have large calculations, low accuracy and success rate, and are difficult to accurately locate in the case of uneven spectrum of lightning signal.
By determining the critical frequency point of the lightning signal, the signal is divided into a new subband based on energy distribution, and the wave reach direction estimation is performed using the rotating signal subspace method to reduce the calculation amount and improve the accuracy.
It realizes accurate estimation of the direction of the lightning signal wave reaching, reduces the calculation amount, improves the accuracy and success rate, and is suitable for lightning signals with uneven spectrum.
Smart Images

Figure CN115061083B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to a method for estimating the direction of arrival of a lightning signal, and specifically relates to a method for realizing lightning signal positioning by improving a general broadband direction of arrival estimation method based on energy distribution within a frequency band. Background Art
[0002] The occurrence of lightning may cause many hazards. The physical effects such as high temperature, violent shock waves and strong electromagnetic radiation generated by lightning can cause huge destructive effects in an instant. It often causes casualties, destroys buildings or communication equipment, causes forest fires, interrupts computer information systems, and causes combustion or even explosions in warehouses, refineries, oil fields, etc., endangering people’s property and personal safety, and posing a great threat to aerospace and other vehicles. Effective positioning methods can provide valuable data for lightning monitoring and meteorological research, and can also provide important support for lightning warning and disaster prevention and mitigation in aerospace systems, power systems and other departments. Lightning positioning is crucial to understanding lightning physics, helping to reveal physical mechanisms and gain an in-depth understanding of the characteristic parameters of lightning discharges, which can help with lightning protection.
[0003] Existing lightning location technologies include traditional magnetic direction finding technology, very low frequency time arrival method, very high frequency interferometer method and positioning technology focusing on array signal processing. Magnetic direction finding technology is easily affected by the terrain of the measurement site or surrounding buildings, and has high requirements for antenna position and antenna installation. It is difficult to meet the requirements in practical applications, and the error is large. The principle of the time arrival method is to use the time difference generated by lightning reaching different measuring stations to locate the lightning position, and the information of a single station has no use value, and the positioning error is large. Interference technology uses multiple cycles of the signal to determine the average time difference within the entire integration period through the phase difference. The interferometer is divided into narrowband interferometer and broadband interferometer. While positioning, the discharge process of lightning can be understood, but the detection distance of interference technology is short and it is difficult to commercialize. It is used more in scientific research work and less in real life. The lightning location method focusing on array signal processing technology eliminates the limitation of the number of antennas on signal processing, and the positioning accuracy can be improved by increasing the number of antennas.
[0004] As the main branch of array signal processing, the principle of direction of arrival estimation is to make the signal incident on the antenna array arranged in a certain pattern, and use a certain method to determine the spatial position of the incident signal source. The direction of arrival estimation method can further improve the positioning accuracy and noise suppression ability by increasing the number of antennas. The traditional broadband MUSIC method uses the orthogonality of the signal and noise subspaces generated by covariance matrix decomposition to construct the signal space spectrum function, and obtains the spatial parameters of the incident signal by searching for peaks. This method has been applied to lightning location and has been proven to have high resolution and strong sidelobe suppression capabilities. The traditional broadband MUSIC method divides the signal into equal-length subsegments in the frequency domain, estimates each subsegment separately, and obtains the average value of the spatial spectrum function. Among them, when dividing the subsegments, only equal-length segments are divided according to the frequency range, without considering the energy distribution in the actual signal band. This segmentation method is more suitable for signals with uniform spectrum. For lightning signals, on the one hand, their bandwidth is wide, and the application of traditional MUSIC methods requires more segments, which is computationally intensive, and the accuracy decreases as the bandwidth increases. On the other hand, the spectrum of lightning signals is unevenly distributed and exponentially distributed, and the idea of uniform segmentation cannot fully consider the energy changes within the frequency band. Inventing a lightning signal direction of arrival estimation method based on energy distribution with less computational effort and higher accuracy has practical application value.
[0005] The Chinese patent "A VHF frequency band cloud-flash lightning detection and positioning system" (application number 201110204820.9) discloses a VHF frequency band cloud-flash lightning detection and positioning system. The system includes a central processing station and multiple detection and processing substations placed in different locations. The central processing station includes a router, a firewall, a communication and monitoring server, a cloud-flash lightning database, a TDOA calculation server and a cloud-flash lightning positioning and solution server. Each detection and processing substation includes a time-frequency unit, a VHF frequency band array antenna, a VHF frequency band array receiving unit, a high-speed synchronous acquisition unit, a VHF frequency band array calibration signal and local oscillator generation unit, a detection and processing substation digital processing subsystem and an embedded controller. The detection and processing substation and the central processing station are connected to the public wide area network by wireless or wired means; through the wide area network, on the one hand, the detection and processing substation uploads the obtained cloud-flash lightning detection data to the central processing station, and on the other hand, the central processing station realizes remote monitoring of the working status of each detection and processing substation. The system can accurately estimate the incident direction of multiple lightning radiation sources occurring simultaneously. The signal arrival direction estimation unit in the digital processing system of the detection and processing substation uses eight lightning waveform data frames from the data cache and control unit, and uses the multiple signal classification method to perform spatial spectrum estimation to obtain the lightning signal arrival direction corresponding to the data frame. This invention provides a complete set of technologies for key links such as cloud-to-cloud lightning signal collection, detection, signal arrival direction estimation, and signal arrival time difference estimation. However, its arrival direction estimation link uses the traditional multiple signal classification method, which does not consider the frequency distribution of lightning signal energy and has low accuracy. Summary of the invention
[0006] The present invention provides a method for estimating the direction of arrival of a lightning signal based on energy distribution, so as to solve the existing problems of large amount of calculation, low precision and low success rate.
[0007] The technical solution adopted by the present invention comprises the following steps:
[0008] Step 1: Determine the critical frequency point of the lightning signal
[0009] According to the US military standard MIL-STD-464, the general description of the lightning signal model is:
[0010] i(t)=I0(e -αt -e -βt )
[0011] Among them, the relevant parameters are I0=218810A, α=11354s -1 , β=647265s -1 ;
[0012] Perform Fourier transform on the lightning signal that meets the above model, analyze it, and take the frequency point where the energy accumulates to 90% as the critical frequency point fh ; Critical frequency point f h The expression that needs to be satisfied is:
[0013]
[0014] Among them, V(j2πf) is the Fourier transform of the lightning signal, j is the imaginary unit, f is the frequency point in the lightning signal spectrum, and the hth frequency point f h It is the critical frequency point where the spectrum energy of lightning signal accumulates to 90%;
[0015] Step 2: Obtain array receiving data generated after the lightning signal passes through the uniform linear array;
[0016] Step 3: The lightning signal array receives data X(f g ) is divided into several new sub-bands;
[0017] (I) Based on the energy distribution, determine the length and number of new sub-bands before and after the critical frequency point.
[0018] (ii) The lightning signal array receives data X(f g ) is divided into several new sub-bands, and the calculation method is:
[0019] W=W1+W2
[0020] Step 4: perform broadband DOA estimation on the W new sub-bands;
[0021] Step 5: Average the spatial spectrum functions of each new sub-band and obtain the direction of arrival estimation result of the lightning signal through spectrum peak search.
[0022] In step (2) of the present invention, P lightning signals conforming to the model described in step 1 are incident on a uniform linear array composed of M omnidirectional array elements, where the array element spacing of the uniform linear array is d;
[0023] Set the first array element as the reference array element, then the time domain array receiving data of the mth array element of the lightning signal is:
[0024]
[0025] Among them, i p (t) is the incident p-th lightning signal, n m (t) is the Gaussian noise added to the mth array element at time t, and the time difference t between the mth array element receiving the pth lightning signal and the first reference array element receiving the pth lightning signal mp is (m-1)dsinθ p / c;
[0026] Perform Fourier transform on the lightning signal time domain array receiving data, and then the lightning signal frequency domain array receiving data is:
[0027]
[0028] Among them, f g is the gth frequency point in the lightning signal spectrum, θ p is the incident angle of the pth lightning signal, a m (θ p ,f g ) is the steering vector of the pth lightning signal, I p (f g ) The pth lightning signal i p The Fourier transform of (t), N m (f g ) is Gaussian noise n m Fourier transform of (t);
[0029] Write the frequency domain array receiving data of lightning signals into matrix form:
[0030] X(f g )=A(θ,f g )I(f g )+N(f g )
[0031] Expand the above matrix into:
[0032] A(θ,f g )=[a(θ1,f g ),a(θ2,f g ),...,a(θ P ,f g )]
[0033]
[0034] I(f g )=[I1(f g ),I2(f g ),...,I P (f g )] T
[0035] N(f g )=[N1(f g ),N2(f g ),...,N M (f g )] T
[0036] X(f g)=[X1(f g ),X2(f g ),...,X M (f g )] T .
[0037] The determination method in step 3 (a) of the present invention is as follows:
[0038] Firstly, according to the lightning signal frequency range, the lightning signal frequency band before and after the critical frequency point is evenly divided to determine the number of equal-length sub-bands before and after the critical frequency point;
[0039] Secondly, each even-numbered subband before the critical frequency point is regarded as the original even-numbered subband before the critical frequency point. According to the ratio of the energy of each original even-numbered subband to the total energy thereof, the length of each new subband before the critical frequency point is determined. The number of new subbands before the critical frequency point is the same as the number of original even-numbered subbands before the critical frequency point. In this way, the number and length of new subbands before the critical frequency point are determined. After the critical frequency point, each even-numbered subband is directly selected as the new subband after the critical frequency point. In this way, the number and length of new subbands after the critical frequency point are determined.
[0040] The lightning signal array received data is divided into equal lengths before and after the critical frequency point, and the original even sub-bands generated are used to determine the new sub-bands before the critical frequency point according to the proportion of the energy of each original even sub-band in its total energy. After the critical point, each even sub-band is directly selected as the new sub-band after the critical frequency point.
[0041] The method for determining the length and number of new sub-bands before the critical frequency point of the lightning signal of the present invention is as follows:
[0042] (1) By analyzing the spectrum of the lightning signal, the frequency band before the critical frequency point is divided into equal-length sub-bands. The number of equal-length sub-bands is Q. Each even-numbered sub-band before the critical frequency point is selected as the original even-numbered sub-band before the critical frequency point. The number of original even-numbered sub-bands before the critical frequency point is Q1.
[0043] Among them, the calculation method of the number Q1 of the original even sub-bands before the critical frequency point of the lightning signal is:
[0044] n is a positive integer
[0045] (2) The length of the frequency band before the critical frequency point of the lightning signal is L. The calculation method of the sum of the sub-band lengths L1 of Q1 original even sub-bands is:
[0046] L1=L / Q×Q1
[0047] Among them, L is the frequency band length before the critical frequency point of the lightning signal, Q1 is the number of original even sub-bands before the critical frequency point of the lightning signal, and Q is the number of equal-length sub-bands divided before the critical frequency point of the lightning signal;
[0048] (3) The energy E of the lightning signal in the qth original even subband before the critical frequency point q The calculation method is:
[0049]
[0050] Among them, V(j2πf r ) is the Fourier transform of the lightning signal, j is the imaginary unit, f r is the rth frequency point in the lightning signal spectrum diagram, the starting frequency point of the original even sub-band whose energy needs to be calculated is the r1th frequency point in the lightning signal spectrum, and the ending frequency point of the original even sub-band whose energy needs to be calculated is the r2th frequency point in the lightning signal spectrum;
[0051] After calculation, the energy of each original even sub-band before the critical frequency point of the lightning signal is determined as
[0052] (4) Sum the energies of the Q1 original even sub-bands before the critical frequency point of the lightning signal calculated above to obtain the total energy E. The calculation method is:
[0053]
[0054] Among them, E q is the energy of the qth original even sub-band of the lightning signal before the critical frequency point, Q1 is the number of original even sub-bands of the lightning signal before the critical frequency point;
[0055] (5) The length C of the qth new subband of the lightning signal before the critical frequency point q The calculation method is:
[0056] C q =E q / E×L1
[0057] Among them, E q is the energy of the qth original even sub-band before the critical frequency point of the lightning signal, E is the sum of the energies of the Q1 original even sub-bands before the critical frequency point of the lightning signal, and L1 is the sum of the sub-band lengths of the Q1 original even sub-bands;
[0058] (6) The calculation method of the number of new sub-bands W1 before the critical frequency point of the lightning signal is as follows:
[0059] W1=Q1
[0060] The number of new subbands is the same as the number of original even subbands.
[0061] The method for determining the length and number of new sub-bands after the critical frequency point of the lightning signal of the present invention is as follows:
[0062] (1) By analyzing the spectrum of the lightning signal, the frequency band after the critical frequency point is divided into equal-length sub-bands. The number of equal-length sub-bands is G. Each even-numbered sub-band after the critical frequency point is selected as the original even-numbered sub-band after the critical frequency point. The number of original even-numbered sub-bands after the critical frequency point is G1.
[0063] Among them, the calculation method of the number of original even sub-bands G1 after the critical frequency point of the lightning signal is:
[0064] n is a positive integer
[0065] (2) The length of the frequency band after the critical frequency point of the lightning signal is L';
[0066] (3) The length B of the kth new subband of the lightning signal after the critical frequency point k The calculation method is as follows:
[0067] B k =L' / G
[0068] Wherein, L' is the frequency band length of the lightning signal after the critical frequency point, and G is the number of equal-length sub-bands into which the lightning signal is divided after the critical frequency point;
[0069] (4) The calculation method of the number of new sub-bands W2 after the critical frequency point of the lightning signal is as follows:
[0070] W2=G1
[0071] The number of new subbands after the critical frequency point is the same as the number of original even subbands after the critical frequency point.
[0072] In step 3, (ii) of the present invention, according to the above principle of dividing new sub-bands based on energy distribution, the critical frequency point of lightning signal is f h , at the critical frequency point f h Before the lightning signal array receives data X(f g ) is divided into W1 new subbands, and the length of the qth new subband is C q , the center point of the original even subband before the critical frequency point is used as the center point of the new subband before the critical frequency point; at the critical frequency point f h After that, the lightning signal array receives the data X(f g ) is divided into W2 new subbands, and the length of the kth new subband is B k , the center point of the original even sub-band after the critical frequency point is used as the center point of the new sub-band after the critical frequency point, so that the lightning signal array receiving data X(f g) is accurately divided to divide the lightning array receiving data into W new sub-bands.
[0073] In step 4 of the present invention, the rotated signal subspace RSS method is used for the divided new subbands, and the specific steps are:
[0074] (1) The incident angle of the lightning signal is estimated in advance using the conventional beamforming method CBF to obtain the estimated angle η;
[0075] (2) Divide the array received data of each new sub-band into K segments in the time domain, perform Fourier transform on each new sub-band, and obtain the array received data of the new sub-band in the frequency domain;
[0076] (3) The array received data of the new subband is divided into N shorter subbands in the frequency domain, and the covariance matrix of the array received data of each shorter subband is calculated as:
[0077] R(f n ,η)=E[X(f n )X H (f n )]
[0078] Where η is the pre-estimated lightning signal incident angle, f n is the nth frequency point in the shorter subband, X(f n ) is the lightning signal at the frequency point f n The array at receives data;
[0079] (4) Determine the reference frequency as f c , reconstruct the array flow matrix A(f n ,η) and A(f c ,η)
[0080] (5) Focusing matrix T(f n ) is determined using the following expression:
[0081] T(f n )=V(f n ) H (f n )
[0082] Among them, U(f n ) and V(f n ) are A(f n ,η)A H (f c ,η), (·) H stands for conjugate transpose;
[0083] (6) The array receiving data at each frequency point is "focused" to the reference frequency f through focusing operationc On the top, the covariance matrix R(f c ,η) is:
[0084] T(f n )R(f n ,η)T H (f n )≈R(f c ,η)
[0085] (7) The covariance matrix R(f c ,η) to perform eigenvalue decomposition and obtain its spatial spectrum function. The spatial spectrum function of the w-th new subband The calculation method is as follows:
[0086]
[0087] in, The value is taken between [0,180] with an interval of 0.01 to achieve the search between [0,180], f c is the reference frequency, a(f c ) is the steering vector of the signal subspace, Z N is the noise subspace.
[0088] The steps of estimating the incident angle of the lightning signal in (1) of the present invention are as follows:
[0089] a. The incident lightning signal passes through a uniform linear array, ignoring the noise, and the received data is:
[0090] II(t)=a(θ)i(t)
[0091] Among them, a(θ) is the guiding vector of the lightning signal, and i(t) is the lightning signal;
[0092] b. Perform phase compensation on the received signal and sum it up to obtain the output signal y(t) of the lightning signal received data after beamforming:
[0093] y(t)=γ H II(t)
[0094] Where γ is the phase compensation vector, (·) H stands for conjugate transpose;
[0095] c. Output power P of conventional beamformer CBF for:
[0096] P CBF =γ H ·R·γ
[0097] Where R is the covariance matrix of the data II(t) received after the lightning signal passes through the uniform linear array;
[0098] d. Search for the peak value of the output power. The angle η corresponding to the peak value is the estimated value of the incident angle of the lightning signal.
[0099] In step 5 of the present invention, the spatial spectrum function of each new sub-band is averaged, and the spatial spectrum function of the lightning signal is The calculation is as follows:
[0100]
[0101] Where W is the number of new subbands, is the spatial spectrum function of the wth new subband in the lightning signal spectrum;
[0102] Spatial spectrum function of lightning signal Perform peak search and find The angle corresponding to the maximum value This is the estimation result of the lightning signal arrival direction.
[0103] The advantages of the present invention are: the present invention fully considers the distribution of energy in the spectrum of lightning signals, determines the frequency point where the energy of the lightning signal spectrum accumulates to 90% as the critical frequency point, calculates the energy of the original even-numbered sub-bands before the critical frequency point, and determines the length of the new sub-band according to the ratio of the energy of each original even-numbered sub-band to its total energy. The higher the energy of the original even-numbered sub-bands, the longer the new sub-bands, and the lower the energy, the shorter the new sub-bands; after the critical frequency point, due to the low energy, the original even-numbered sub-bands are directly selected as new sub-bands for processing. By estimating the direction of arrival of each new sub-band, the accurate estimation of the direction of arrival of the lightning signal is achieved. Compared with the traditional broadband MUSIC method, the data processed by the present invention is reduced by half, and the amount of calculation is also reduced by half. Compared with the traditional broadband MUSIC method, its accuracy and success rate are also improved. The present invention has better lightning signal direction of arrival estimation performance. BRIEF DESCRIPTION OF THE DRAWINGS
[0104] Figure 1 This is the far-field one-dimensional linear array DOA estimation geometric model diagram;
[0105] Figure 2 is a flow chart of the present invention;
[0106] Figure 3 It is a schematic diagram of the original even sub-bands generated after the lightning signal is divided into equal lengths before and after the critical frequency point;
[0107] Figure 4 It is a schematic diagram of the new sub-band determined by the energy of the lightning signal before and after the critical frequency point;
[0108] Figure 5 This is a comparison diagram of the root mean square error of the lightning signal arrival direction estimation using the method of the present invention and the traditional method;
[0109] Figure 6 This is a comparison chart of the success rates of estimating the direction of arrival of lightning signals using the method of the present invention and the traditional method. DETAILED DESCRIPTION
[0110] Combine the following Figure 2 The present invention is described in detail with reference to the flow chart.
[0111] Step 1: Determine the critical frequency point of the lightning signal
[0112] According to the US military standard MIL-STD-464, the general description of the lightning signal model is:
[0113] i(t)=I0(e -αt -e -βt )
[0114] Among them, the relevant parameters are I0=218810A, α=11354s -1 , β=647265s -1 ;
[0115] Perform Fourier transform on the lightning signal that conforms to the above model, and we get Figure 3 The spectrum of the lightning signal shown in the curve is analyzed, and the frequency point where the energy accumulates to 90% is taken as the critical frequency point f h ;
[0116] Figure 3 Medium h That is the critical frequency point, the critical frequency point f h The expression that needs to be satisfied is:
[0117]
[0118] Among them, V(j2πf) is the Fourier transform of the lightning signal, j is the imaginary unit, f is the frequency point in the lightning signal spectrum, and the hth frequency point f h It is the critical frequency point where the spectrum energy of lightning signal accumulates to 90%;
[0119] Step 2: Obtain the array receiving data generated after the lightning signal passes through the uniform linear array
[0120] P lightning signals that meet the model description of step 1 are incident on a uniform linear array composed of M omnidirectional array elements, and the array element spacing of the uniform linear array is d;
[0121] Set the first array element as the reference array element, then the time domain array receiving data of the mth array element of the lightning signal is:
[0122]
[0123] Among them, i p (t) is the incident p-th lightning signal, n m (t) is the Gaussian noise added to the mth array element at time t, and the time difference t between the mth array element receiving the pth lightning signal and the first reference array element receiving the pth lightning signal mp is (m-1)dsinθ p / c;
[0124] Perform Fourier transform on the lightning signal time domain array receiving data, and then the lightning signal frequency domain array receiving data is:
[0125]
[0126] Among them, f g is the gth frequency point in the lightning signal spectrum, θ p is the incident angle of the pth lightning signal, a m (θ p ,f g ) is the steering vector of the pth lightning signal, I p (f g ) The pth lightning signal i p The Fourier transform of (t), N m (f g ) is Gaussian noise n m Fourier transform of (t);
[0127] Write the frequency domain array receiving data of lightning signals into matrix form:
[0128] X(f g )=A(θ,f g )I(f g )+N(f g )
[0129] Expand the above matrix into:
[0130] A(θ,f g )=[a(θ1,f g ),a(θ2,f g ),...,a(θ P ,f g )]
[0131]
[0132] I(fg )=[I1(f g ),I2(f g ),...,I P (f g )] T
[0133] N(f g )=[N1(f g ),N2(f g ),...,N M (f g )] T
[0134] X(f g )=[X1(f g ),X2(f g ),...,X M (f g )] T
[0135] Step 3: The lightning signal array receives data X(f g ) is divided into several new sub-bands
[0136] (I) Based on the energy distribution, the length and number of new sub-bands before and after the critical frequency point are determined as follows:
[0137] Firstly, according to the lightning signal frequency range, the lightning signal frequency band before and after the critical frequency point is evenly divided to determine the number of equal-length sub-bands before and after the critical frequency point;
[0138] Secondly, each even-numbered subband before the critical frequency point is regarded as the original even-numbered subband before the critical frequency point. According to the ratio of the energy of each original even-numbered subband to the total energy thereof, the length of each new subband before the critical frequency point is determined. The number of new subbands before the critical frequency point is the same as the number of original even-numbered subbands before the critical frequency point. In this way, the number and length of new subbands before the critical frequency point are determined. After the critical frequency point, each even-numbered subband is directly selected as the new subband after the critical frequency point. In this way, the number and length of new subbands after the critical frequency point are determined.
[0139] The lightning signal array received data is divided into equal lengths before and after the critical frequency point, and the original even sub-bands generated are as follows: Figure 3 As shown in the shaded part, before the critical frequency point, the new sub-band before the critical frequency point is determined according to the proportion of the energy of each original even sub-band in its total energy. After the critical point, each even sub-band is directly selected as the new sub-band after the critical frequency point. The new sub-bands for the lightning signal array receiving data are divided as follows: Figure 4 The shaded area is shown;
[0140] Before the critical frequency point of the lightning signal, the length and number of the new sub-band are determined as follows:
[0141] (1) By analyzing the spectrum of the lightning signal, the frequency band before the critical frequency point is divided into equal-length sub-bands. The number of equal-length sub-bands is Q. Each even-numbered sub-band before the critical frequency point is selected as the original even-numbered sub-band before the critical frequency point. The number of original even-numbered sub-bands before the critical frequency point is Q1.
[0142] Among them, the calculation method of the number Q1 of the original even sub-bands before the critical frequency point of the lightning signal is:
[0143] n is a positive integer
[0144] (2) The length of the frequency band before the critical frequency point of the lightning signal is L. The calculation method of the sum of the sub-band lengths L1 of Q1 original even sub-bands is:
[0145] L1=L / Q×Q1
[0146] Among them, L is the frequency band length before the critical frequency point of the lightning signal, Q1 is the number of original even sub-bands before the critical frequency point of the lightning signal, and Q is the number of equal-length sub-bands divided before the critical frequency point of the lightning signal;
[0147] (3) The energy E of the lightning signal in the qth original even subband before the critical frequency point q The calculation method is:
[0148]
[0149] Among them, V(j2πf r ) is the Fourier transform of the lightning signal, j is the imaginary unit, f r is the rth frequency point in the lightning signal spectrum diagram, the starting frequency point of the original even sub-band whose energy needs to be calculated is the r1th frequency point in the lightning signal spectrum, and the ending frequency point of the original even sub-band whose energy needs to be calculated is the r2th frequency point in the lightning signal spectrum;
[0150] After calculation, the energy of each original even sub-band before the critical frequency point of the lightning signal is determined as
[0151] (4) Sum the energies of the Q1 original even sub-bands before the critical frequency point of the lightning signal calculated above to obtain the total energy E. The calculation method is:
[0152]
[0153] Among them, E q is the energy of the qth original even sub-band of the lightning signal before the critical frequency point, Q1 is the number of original even sub-bands of the lightning signal before the critical frequency point;
[0154] (5) The length C of the qth new subband of the lightning signal before the critical frequency point q The calculation method is:
[0155] C q =E q / E×L1
[0156] Among them, E q is the energy of the qth original even sub-band before the critical frequency point of the lightning signal, E is the sum of the energies of the Q1 original even sub-bands before the critical frequency point of the lightning signal, and L1 is the sum of the sub-band lengths of the Q1 original even sub-bands;
[0157] (6) The calculation method of the number of new sub-bands W1 before the critical frequency point of the lightning signal is as follows:
[0158] W1=Q1
[0159] The number of new subbands is the same as the number of original even subbands;
[0160] After the lightning signal reaches the critical frequency point, the length and number of the new sub-band are determined as follows:
[0161] (1) By analyzing the spectrum of the lightning signal, the frequency band after the critical frequency point is divided into equal-length sub-bands. The number of equal-length sub-bands is G. Each even-numbered sub-band after the critical frequency point is selected as the original even-numbered sub-band after the critical frequency point. The number of original even-numbered sub-bands after the critical frequency point is G1.
[0162] Among them, the calculation method of the number of original even sub-bands G1 after the critical frequency point of the lightning signal is:
[0163] n is a positive integer
[0164] (2) The length of the frequency band after the critical frequency point of the lightning signal is L';
[0165] (3) The length B of the kth new subband of the lightning signal after the critical frequency point k The calculation method is as follows:
[0166] B k =L' / G
[0167] Wherein, L' is the frequency band length of the lightning signal after the critical frequency point, and G is the number of equal-length sub-bands into which the lightning signal is divided after the critical frequency point;
[0168] (4) The calculation method of the number of new sub-bands W2 after the critical frequency point of the lightning signal is as follows:
[0169] W2=G1
[0170] The number of new subbands after the critical frequency point is the same as the number of original even subbands after the critical frequency point;
[0171] (ii) The lightning signal array receives data X(f g ) is divided into several new sub-bands
[0172] According to the above principle of dividing new sub-bands based on energy distribution, the critical frequency point of lightning signal is f h , at the critical frequency point f h Before the lightning signal array receives data X(f g ) is divided into W1 new subbands, and the length of the qth new subband is C q , the center point of the original even subband before the critical frequency point is used as the center point of the new subband before the critical frequency point; at the critical frequency point f h After that, the lightning signal array receives the data X(f g ) is divided into W2 new subbands, and the length of the kth new subband is B k , the center point of the original even sub-band after the critical frequency point is used as the center point of the new sub-band after the critical frequency point, so that the lightning signal array receiving data X(f g ) is accurately divided to divide the lightning array receiving data into W new sub-bands. The calculation method is:
[0173] W=W1+W2
[0174] Step 4: Perform broadband DOA estimation on the W new sub-bands
[0175] The rotated signal subspace RSS method is used for the newly divided subbands. The specific steps are:
[0176] (1) The incident angle of the lightning signal is estimated in advance using the conventional beamforming (CBF) method to obtain the estimated angle η;
[0177] The steps for estimating the lightning signal incident angle are as follows:
[0178] a. The incident lightning signal passes through a uniform linear array, ignoring the noise, and the received data is:
[0179] II(t)=a(θ)i(t)
[0180] Among them, a(θ) is the guiding vector of the lightning signal, and i(t) is the lightning signal;
[0181] b. Perform phase compensation on the received signal and sum it up to obtain the output signal y(t) of the lightning signal received data after beamforming:
[0182] y(t)=γ HII(t)
[0183] Where γ is the phase compensation vector, (·) H stands for conjugate transpose;
[0184] c. Output power P of conventional beamformer CBF for:
[0185] P CBF =γ H ·R·γ
[0186] Where R is the covariance matrix of the data II(t) received after the lightning signal passes through the uniform linear array;
[0187] d. Search for the peak value of the output power. The angle η corresponding to the peak value is the estimated value of the incident angle of the lightning signal;
[0188] (2) Divide the array received data of each new sub-band into K segments in the time domain, perform Fourier transform on each new sub-band, and obtain the array received data of the new sub-band in the frequency domain;
[0189] (3) The array received data of the new subband is divided into N shorter subbands in the frequency domain, and the covariance matrix of the array received data of each shorter subband is calculated as:
[0190] R(f n ,η)=E[X(f n )X H (f n )]
[0191] Where η is the pre-estimated lightning signal incident angle, f n is the nth frequency point in the shorter subband, X(f n ) is the lightning signal at the frequency point f n The array at receives the data.
[0192] (4) Determine the reference frequency as f c , reconstruct the array flow matrix A(f n ,η) and A(f c ,η)
[0193] (5) Focusing matrix T(f n ) is determined using the following expression:
[0194] T(f n )=V(f n ) H (f n )
[0195] Among them, U(f n ) and V(f n) are A(f n ,η)A H (f c ,η), (·) H stands for conjugate transpose.
[0196] (6) The array receiving data at each frequency point is "focused" to the reference frequency f through focusing operation c On the top, the covariance matrix R(f c ,η) is:
[0197] T(f n )R(f n ,η)T H (f n )≈R(f c ,η)
[0198] (7) The covariance matrix R(f c ,η) to perform eigenvalue decomposition and obtain its spatial spectrum function. The spatial spectrum function of the w-th new subband The calculation method is as follows:
[0199]
[0200] in, The value is taken between [0,180] with an interval of 0.01 to achieve the search between [0,180], f c is the reference frequency, a(f c ) is the steering vector of the signal subspace, Z N is the noise subspace;
[0201] Step 5: Average the spatial spectrum functions of each new sub-band and obtain the direction of arrival estimation result of the lightning signal through spectrum peak search;
[0202] The spatial spectrum function of each new sub-band is averaged, and the spatial spectrum function of the lightning signal is The calculation is as follows:
[0203]
[0204] Where W is the number of new subbands, is the spatial spectrum function of the wth new subband in the lightning signal spectrum;
[0205] Spatial spectrum function of lightning signal Perform peak search and find The angle corresponding to the maximum value This is the estimated result of the lightning signal arrival direction.
[0206] The present invention fully considers the energy change in the spectrum of lightning signals. According to the characteristics of lightning signals, the frequency point where the energy of the lightning signal spectrum accumulates 90% is taken as the critical frequency point, and the signal is processed differently before and after the critical frequency point. Before the critical frequency point, the energy of the original even-numbered sub-band is calculated, and the length of the new sub-band is determined according to the ratio of the energy of each original even-numbered sub-band to its total energy. The higher the energy of the original even-numbered sub-band, the longer the new sub-band, and the lower the energy, the shorter the new sub-band; after the critical frequency point, due to the low energy, the original even-numbered sub-band is directly selected as the new sub-band for processing. A general broadband direction of arrival estimation method is used for each new sub-band, and the average of the obtained results is taken to obtain the direction of arrival estimation result of the lightning signal.
[0207] Compared with the existing method for estimating the direction of arrival of lightning signals, the present invention greatly reduces the amount of calculation and improves the accuracy and success rate.
[0208] The experiment uses an i7-7500U processor with 8.00GB main memory and Matlab 2016a as the simulation platform. In order to ensure the accuracy of the experiment, 200 Monte Carlo experiments are performed.
[0209] Figure 5 The root mean square error between the estimated angle and the true angle obtained by estimating the direction of arrival of lightning signals using the method of the present invention and the traditional method after 200 Monte Carlo experiments within a certain signal-to-noise ratio range. Figure 5 As shown, under low signal-to-noise ratio, the accuracy of the traditional method is slightly higher than that of the method of the present invention, but as the signal-to-noise ratio increases, the accuracy of the method of the present invention is significantly improved.
[0210] Figure 6 After 200 Monte Carlo experiments, within a certain signal-to-noise ratio range, the method of the present invention and the traditional method are used to estimate the direction of arrival of lightning signals, and the success rate of the estimation is compared. Figure 6 As shown, the success rate of the method of the present invention is higher than that of the traditional method.
[0211] It is experimentally measured that the time required for estimating the arrival direction of two lightning signals using the traditional method is 25.87 seconds, while the time required for estimating the arrival direction of two lightning signals using the method of the present invention is 12.48 seconds, and the operation time is greatly reduced.
[0212] In summary, the method of the present invention not only improves the accuracy of estimation, but also improves the success rate of estimation and reduces the amount of calculation.
[0213] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for estimating the direction of arrival of a lightning signal based on energy distribution, characterized in that: The following steps are involved: Step 1: Determine the critical frequency point of the lightning signal According to the US military standard MIL-STD-464, the general description of the lightning signal model is: i(t)=I0(e -αt -Alright -βt ) Among them, the relevant parameters are I0=218810A, α=11354s -1 , β=647265s -1 ; Perform Fourier transform on the lightning signal that meets the above model, analyze it, and take the frequency point where the energy accumulates to 90% as the critical frequency point f h ; Critical frequency point f h The expression that needs to be satisfied is: Among them, V(j2πf) is the Fourier transform of the lightning signal, j is the imaginary unit, f is the frequency point in the lightning signal spectrum, and the hth frequency point f h It is the critical frequency point where the spectrum energy of lightning signal accumulates to 90%; Step 2: Obtain array receiving data generated after the lightning signal passes through the uniform linear array; Step 3: The lightning signal array receives data X(f g ) is divided into several new sub-bands (I) Based on the energy distribution, determine the length and number of new sub-bands before and after the critical frequency point. (ii) The lightning signal array receives data X(f g ) is divided into several new sub-bands, and the calculation method is: W=W1+W2 Step 4: perform broadband DOA estimation on the W new sub-bands; Step 5: Average the spatial spectrum functions of each new sub-band and obtain the direction of arrival estimation result of the lightning signal through spectrum peak search.
2. The method for estimating the direction of arrival of a lightning signal based on energy distribution according to claim 1, characterized in that: In the step (2), P lightning signals conforming to the model described in step 1 are incident on a uniform linear array composed of M omnidirectional array elements, where the array element spacing of the uniform linear array is d; Set the first array element as the reference array element, then the time domain array receiving data of the mth array element of the lightning signal is: Among them, i p (t) is the incident p-th lightning signal, n m (t) is the Gaussian noise added to the mth array element at time t, and the time difference t between the mth array element receiving the pth lightning signal and the first reference array element receiving the pth lightning signal mp is (m-1)dsinθ p / c; Perform Fourier transform on the lightning signal time domain array receiving data, and then the lightning signal frequency domain array receiving data is: Among them, f g is the gth frequency point in the lightning signal spectrum, θ p is the incident angle of the pth lightning signal, a m (θ p ,f g ) is the steering vector of the pth lightning signal, I p (f g ) The pth lightning signal i p The Fourier transform of (t), N m (f g ) is Gaussian noise n m Fourier transform of (t); Write the frequency domain array receiving data of lightning signals into matrix form: X(f g )=A(θ,f g )I(f g )+N(f g ) Expand the above matrix into: A(θ,f g )=[a(θ1,f g ),a(θ2,f g ),...,a(θ P ,f g )] I(f g )=[I1(f g ),I2(f g ),...,I P (f g )] T N(f g )=[N1(f g ),N2(f g ),...,N M (f g )] T X(f g )=[X1(f g ),X2(f g ),...,X M (f g )] T 。 3. The method for estimating the direction of arrival of a lightning signal based on energy distribution according to claim 1, characterized in that: The determination method in step 3 (a) is as follows: Firstly, according to the lightning signal frequency range, the lightning signal frequency band before and after the critical frequency point is evenly divided to determine the number of equal-length sub-bands before and after the critical frequency point; Secondly, each even-numbered subband before the critical frequency point is regarded as the original even-numbered subband before the critical frequency point. According to the ratio of the energy of each original even-numbered subband to the total energy thereof, the length of each new subband before the critical frequency point is determined. The number of new subbands before the critical frequency point is the same as the number of original even-numbered subbands before the critical frequency point. In this way, the number and length of new subbands before the critical frequency point are determined. After the critical frequency point, each even-numbered subband is directly selected as the new subband after the critical frequency point. In this way, the number and length of new subbands after the critical frequency point are determined. The lightning signal array received data is divided into equal lengths before and after the critical frequency point, and the original even sub-bands generated are used to determine the new sub-bands before the critical frequency point according to the proportion of the energy of each original even sub-band in its total energy. After the critical point, each even sub-band is directly selected as the new sub-band after the critical frequency point.
4. The method for estimating the direction of arrival of a lightning signal based on energy distribution according to claim 3, characterized in that: Before the critical frequency point of the lightning signal, the length and number of the new sub-band are determined as follows: (1) By analyzing the spectrum of the lightning signal, the frequency band before the critical frequency point is divided into equal-length sub-bands. The number of equal-length sub-bands is Q. Each even-numbered sub-band before the critical frequency point is selected as the original even-numbered sub-band before the critical frequency point. The number of original even-numbered sub-bands before the critical frequency point is Q1. Among them, the calculation method of the number Q1 of the original even sub-bands before the critical frequency point of the lightning signal is: n is a positive integer (2) The length of the frequency band before the critical frequency point of the lightning signal is L. The calculation method of the sum of the sub-band lengths L1 of Q1 original even sub-bands is: L1=L / Q×Q1 Among them, L is the frequency band length before the critical frequency point of the lightning signal, Q1 is the number of original even sub-bands before the critical frequency point of the lightning signal, and Q is the number of equal-length sub-bands divided before the critical frequency point of the lightning signal; (3) The energy E of the lightning signal in the qth original even subband before the critical frequency point q The calculation method is: Among them, V(j2πf r ) is the Fourier transform of the lightning signal, j is the imaginary unit, f r is the rth frequency point in the lightning signal spectrum diagram, the starting frequency point of the original even sub-band whose energy needs to be calculated is the r1th frequency point in the lightning signal spectrum, and the ending frequency point of the original even sub-band whose energy needs to be calculated is the r2th frequency point in the lightning signal spectrum; After calculation, the energy of each original even sub-band before the critical frequency point of the lightning signal is determined as (4) Sum the energies of the Q1 original even sub-bands before the critical frequency point of the lightning signal calculated above to obtain the total energy E. The calculation method is: Among them, E q is the energy of the qth original even sub-band of the lightning signal before the critical frequency point, Q1 is the number of original even sub-bands of the lightning signal before the critical frequency point; (5) The length C of the qth new subband of the lightning signal before the critical frequency point q The calculation method is: C q =And q / E×L1 Among them, E q is the energy of the qth original even sub-band before the critical frequency point of the lightning signal, E is the sum of the energies of the Q1 original even sub-bands before the critical frequency point of the lightning signal, and L1 is the sum of the sub-band lengths of the Q1 original even sub-bands; (6) The calculation method of the number of new sub-bands W1 before the critical frequency point of the lightning signal is as follows: W1=Q1 The number of new subbands is the same as the number of original even subbands.
5. The method for estimating the direction of arrival of a lightning signal based on energy distribution according to claim 3, characterized in that: After the lightning signal reaches the critical frequency point, the length and number of the new sub-band are determined as follows: (1) By analyzing the spectrum of the lightning signal, the frequency band after the critical frequency point is divided into equal-length sub-bands. The number of equal-length sub-bands is G. Each even-numbered sub-band after the critical frequency point is selected as the original even-numbered sub-band after the critical frequency point. The number of original even-numbered sub-bands after the critical frequency point is G1. Among them, the calculation method of the number of original even sub-bands G1 after the critical frequency point of the lightning signal is: n is a positive integer (2) The length of the frequency band after the critical frequency point of the lightning signal is L'; (3) The length B of the kth new subband of the lightning signal after the critical frequency point k The calculation method is as follows: B k =L' / G Wherein, L' is the frequency band length of the lightning signal after the critical frequency point, and G is the number of equal-length sub-bands into which the lightning signal is divided after the critical frequency point; (4) The calculation method of the number of new sub-bands W2 after the critical frequency point of the lightning signal is as follows: W2=G1 The number of new subbands after the critical frequency point is the same as the number of original even subbands after the critical frequency point.
6. The method for estimating the direction of arrival of a lightning signal based on energy distribution according to claim 1, characterized in that: In step 3, (ii), according to the above principle of dividing new sub-bands based on energy distribution, the critical frequency point of lightning signal is f h , at the critical frequency point f h Before the lightning signal array receives data X(f g ) is divided into W1 new subbands, and the length of the qth new subband is C q , the center point of the original even subband before the critical frequency point is used as the center point of the new subband before the critical frequency point; at the critical frequency point f h After that, the lightning signal array receives the data X(f g ) is divided into W2 new subbands, and the length of the kth new subband is B k , the center point of the original even sub-band after the critical frequency point is used as the center point of the new sub-band after the critical frequency point, so that the lightning signal array receiving data X(f g ) is accurately divided to divide the lightning array receiving data into W new sub-bands.
7. The method for estimating the direction of arrival of a lightning signal based on energy distribution according to claim 1, characterized in that: In step 4, the rotated signal subspace RSS method is used for the divided new subbands, and the specific steps are: (1) The incident angle of the lightning signal is estimated in advance using the conventional beamforming method CBF to obtain the estimated angle η; (2) Divide the array received data of each new sub-band into K segments in the time domain, perform Fourier transform on each new sub-band, and obtain the array received data of the new sub-band in the frequency domain; (3) The array received data of the new subband is divided into N shorter subbands in the frequency domain, and the covariance matrix of the array received data of each shorter subband is calculated as: R(f n ,η)=E[X(f n )X H (f n )] Where η is the pre-estimated lightning signal incident angle, f n is the nth frequency point in the shorter subband, X(f n ) is the lightning signal at the frequency point f n The array at receives data; (4) Determine the reference frequency as f c , reconstruct the array flow matrix A(f n ,η) and A(f c ,η) (5) Focusing matrix T(f n ) is determined using the following expression: T(f n )=V(f n )U H (f n ) Among them, U(f n ) and V(f n ) are A(f n ,η)A H (f c ,η), (·) H stands for conjugate transpose; (6) The array receiving data at each frequency point is "focused" to the reference frequency f through focusing operation c On the top, the covariance matrix R(f c ,η) is: T(f n )R(f n ,η)T H (f n )≈R(f c ,η) (7) The covariance matrix R(f c ,η) to perform eigenvalue decomposition and obtain its spatial spectrum function. The spatial spectrum function of the w-th new subband The calculation method is as follows: in, The value is taken between [0,180] with an interval of 0.01 to achieve the search between [0,180], f c is the reference frequency, a(f c ) is the steering vector of the signal subspace, Z N is the noise subspace.
8. The method for estimating the direction of arrival of lightning signals based on energy distribution according to claim 7, characterized in that: The steps of estimating the lightning signal incident angle in (1) are as follows: a. The incident lightning signal passes through a uniform linear array, ignoring the noise, and the received data is: II(t)=a(θ)i(t) Among them, a(θ) is the guiding vector of the lightning signal, and i(t) is the lightning signal; b. Perform phase compensation on the received signal and sum it up to obtain the output signal y(t) of the lightning signal received data after beamforming: y(t)=γ H ·II(t) Where γ is the phase compensation vector, (·) H stands for conjugate transpose; c. Output power P of conventional beamformer CBF for: P CBF =c H ·R·c Where R is the covariance matrix of the data II(t) received after the lightning signal passes through the uniform linear array; d. Search for the peak value of the output power. The angle η corresponding to the peak value is the estimated value of the incident angle of the lightning signal.
9. The method for estimating the direction of arrival of a lightning signal based on energy distribution according to claim 1, characterized in that: In step 5, the spatial spectrum functions of each new sub-band are averaged, and the spatial spectrum function of the lightning signal is The calculation is as follows: Where W is the number of new subbands, is the spatial spectrum function of the wth new subband in the lightning signal spectrum; Spatial spectrum function of lightning signal Perform peak search and find The angle corresponding to the maximum value This is the estimated result of the lightning signal arrival direction.
