Lightning electromagnetic wave positioning method, system and equipment based on single-station detection

By using deep learning neural network model in a single-station lightning positioning system, combining GPS or Beidou signals, lightning electromagnetic pulse data is collected in real time, and the problem of large positioning error in a single-station is solved, achieving high-precision positioning and real-time detection of lightning events is achieved. , improves the efficiency and adaptability of the system.

CN114814960BActive Publication Date: 2024-11-26INST OF ELECTRICAL ENG CHINESE ACAD OF SCI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210239373.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-11
Publication Date
2024-11-26
Estimated Expiration
2042-03-11

AI Technical Summary

Technical Problem

The existing single-station lightning positioning system has problems such as large positioning error, low efficiency, small detection range, and easy interference, which limits its promotion and use.

Method used

Using a lightning electromagnetic wave positioning method based on single-station detection, data is collected in real time through lightning electromagnetic pulse detection equipment, combined with GPS or Beidou time signal, and a lightning electromagnetic pulse propagation distance prediction model constructed by deep learning neural networks to calculate the propagation distance of lightning electromagnetic pulses and wave orientation to achieve high-precision positioning.

Benefits of technology

It realizes high-precision detection of lightning events within a range of 1,000 kilometers, with positioning accuracy better than 5% of the signal propagation distance, improving positioning accuracy and efficiency, suitable for a variety of computing platforms, with good model portability and real-time detection capabilities .

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114814960B_ABST
    Figure CN114814960B_ABST
Patent Text Reader

Abstract

The present invention belongs to the fields of lightning detection, artificial intelligence and deep learning, and specifically relates to a lightning electromagnetic wave positioning method, system and equipment based on single-station detection, aiming to solve the problems of large positioning error, low efficiency, small range and susceptibility of positioning system to interference in existing single-station lightning positioning methods. The present invention includes: real-time acquisition of lightning electromagnetic pulse waveform data by lightning electromagnetic pulse detection equipment, and demodulation of received GPS or Beidou timing signals into high-precision timing information; calculation of the propagation distance of lightning electromagnetic pulses by trained lightning electromagnetic pulse propagation distance prediction model; calculation of the wave azimuth of lightning electromagnetic pulses based on the collected lightning electromagnetic pulse waveform data; fusion of the propagation distance and wave azimuth of lightning electromagnetic pulses to obtain the positioning result of lightning electromagnetic pulses. The present invention can realize high-precision detection of various types of lightning electromagnetic pulses, with a single-station detection range of more than 1,000 kilometers and a positioning accuracy better than 5%.
Need to check novelty before this filing date? Find Prior Art

Description

Background Art

[0002] Lightning is a high-intensity instantaneous discharge process that occurs frequently in nature. On the one hand, it maintains the electrical balance between the atmospheric ionosphere and the earth; on the other hand, lightning disasters have been listed as "one of the ten most serious natural disasters" by the United Nations Decade Committee for Disaster Reduction. The electromagnetic wave spectrum radiated by lightning is extremely wide and extremely intense, releasing a large amount of electromagnetic energy in the frequency range from less than 1Hz to nearly 300MHz. Therefore, electromagnetic pulse detection technology has become the preferred method for lightning detection.

[0003] At present, the commonly used ground-based lightning electromagnetic pulse detection technology is mainly based on the positioning method of multi-station time difference and direction finding. The mainstream commercial lightning positioning system uses the very low frequency / low frequency (VLF / LF) detection band and requires the use of at least 4 detection stations. The baseline distance between the stations is generally 100-150km to achieve real-time three-dimensional positioning of lightning events in the detection network. The advantage of this system is its high positioning accuracy, and the average positioning error of lightning events occurring in the network is better than 500 meters. The disadvantages are that multi-station network detection is required; each detection station needs to maintain high-precision clock synchronization; the station location is fixed and cannot be applied to mobile systems.

[0004] The single-station lightning location system is an important component of lightning detection. With its simple operation and flexible mobility, the system has broad application prospects in airports, ships, aviation and other fields. At present, the single-station lightning location system mainly uses electromagnetic pulse detection technology to achieve rapid detection of lightning events. Although the single-station lightning location technology has been proposed for many years, there are still many problems such as large positioning error, low detection efficiency, small detection range, and susceptibility to interference, which seriously limit the promotion and use of the single-station lightning location system. Summary of the invention

[0005] In order to solve the above problems in the prior art, namely, the existing single-station lightning positioning method has large positioning error, low efficiency, small range, and the positioning system is susceptible to interference, the present invention provides a lightning electromagnetic wave positioning method based on single-station detection, the positioning method comprises:

[0006] Step S10, collecting lightning electromagnetic pulse waveform data in real time through a lightning electromagnetic pulse detection device, and demodulating the received GPS or Beidou timing signal into high-precision timing information;

[0007] Step S20, based on the collected lightning electromagnetic pulse waveform data and the high-precision timing information, and using a trained lightning electromagnetic pulse propagation distance prediction model to calculate the propagation distance of the lightning electromagnetic pulse;

[0008] Step S30, calculating the wave direction of the lightning electromagnetic pulse based on the collected lightning electromagnetic pulse waveform data;

[0009] Step S40, fusing the lightning electromagnetic pulse propagation distance and wave direction to obtain the lightning electromagnetic pulse positioning result.

[0010] In some preferred embodiments, the lightning electromagnetic pulse propagation distance prediction model is constructed based on a deep learning neural network, including an input layer, a multi-dimensional feature extraction network, a multi-feature fusion and mapping network, and an output layer.

[0011] In some preferred embodiments, the multi-dimensional feature extraction network includes three branch networks with the same structure, which are respectively used to extract the electric field signal, the north-south magnetic field signal and the east-west magnetic field signal of the lightning electromagnetic pulse;

[0012] The branch network includes a first input layer, multiple one-dimensional convolutional neural network computing layers, a batch normalization computing layer, an activation computing layer, a residual block, a maximum pooling computing layer, a global mean pooling computing layer and a first output layer.

[0013] In some preferred embodiments, the first input layer is used to transmit the collected lightning electromagnetic pulse data with a dimension of 1×N, where N is the sampling rate of the lightning electromagnetic pulse detection equipment and the time length of a single sampling;

[0014] The first output layer is used to output feature data with a dimension of 1×F extracted by the three branch networks respectively, where F is the number of features output by each branch network.

[0015] In some preferred embodiments, the multi-feature fusion and mapping network includes a second input layer, a data fusion layer, a residual neural network, a fully connected layer, an activation layer and a second output layer.

[0016] In some preferred embodiments, the second input layer is used to transmit high-precision timing information and feature information of electric field signals, north-south magnetic field signals, and east-west magnetic field signals with dimensions of 1×F respectively output by three branch networks of the multi-dimensional feature extraction network;

[0017] The second output layer is used to output fused data with a dimension of 1×(3×F+1).

[0018] In some preferred embodiments, the lightning electromagnetic pulse propagation distance prediction model is trained by:

[0019] Step B10, obtaining a set amount of lightning electromagnetic pulse waveform data and corresponding high-precision timing information as a model training data set, and using the actual propagation distance of the lightning electromagnetic pulse corresponding to the waveform data as a label of the training data;

[0020] Step B20, selecting any training data, and obtaining the corresponding lightning electromagnetic pulse propagation distance through the constructed lightning electromagnetic pulse propagation distance prediction model;

[0021] Step B30, calculating the loss value between the propagation distance of the lightning electromagnetic pulse and the label of the current training data by using a set loss function;

[0022] Step B40, updating the parameters of the lightning electromagnetic pulse propagation distance prediction model through the optimization function, and performing iterative model training through the training data set until the set training end condition is reached, thereby obtaining a trained lightning electromagnetic pulse propagation distance prediction model.

[0023] In some preferred embodiments, the wave direction of the lightning electromagnetic pulse is calculated by:

[0024]

[0025]

[0026] Among them, θ is the wave direction of the lightning electromagnetic pulse, E is the maximum amplitude of the electric field signal of the lightning electromagnetic pulse, and B NS is the maximum amplitude of the north-south magnetic field signal of the lightning electromagnetic pulse, B EW It is the maximum amplitude of the east-west magnetic field signal of the lightning electromagnetic pulse.

[0027] Another aspect of the present invention provides a lightning electromagnetic wave positioning system based on single-station detection, the positioning system comprising the following modules:

[0028] A signal acquisition module is configured to collect lightning electromagnetic pulse waveform data in real time through a lightning electromagnetic pulse detection device;

[0029] The timing module is configured to receive GPS or Beidou timing signals and output high-precision timing information through the signal demodulation module and pass it to the lightning electromagnetic pulse propagation distance prediction model;

[0030] A distance prediction module is configured to calculate the propagation distance of the lightning electromagnetic pulse based on the collected lightning electromagnetic pulse waveform data and the high-precision timing information through a trained lightning electromagnetic pulse propagation distance prediction model;

[0031] A wave orientation calculation module configured to calculate the wave orientation of the lightning electromagnetic pulse based on the collected lightning electromagnetic pulse waveform data;

[0032] The fusion positioning module is configured to fuse the propagation distance and wave direction of the lightning electromagnetic pulse to obtain the positioning result of the lightning electromagnetic pulse.

[0033] According to a third aspect of the present invention, an electronic device is provided, comprising:

[0034] at least one processor; and

[0035] a memory communicatively connected to at least one of the processors; wherein,

[0036] The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the above-mentioned lightning electromagnetic wave positioning method based on single-station detection.

[0037] In a fourth aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to be executed by the computer to implement the above-mentioned lightning electromagnetic wave positioning method based on single-station detection.

[0038] Beneficial effects of the present invention:

[0039] (1) The lightning electromagnetic wave positioning method based on single-station detection in the present invention combines the calculation of the electromagnetic pulse arrival azimuth and the propagation distance prediction model to achieve high-precision detection of lightning events occurring within a range of 1,000 kilometers using a single detection station, with a positioning accuracy better than 5% of the signal propagation distance.

[0040] (2) The lightning electromagnetic wave locating method of the present invention is based on single-station detection. The lightning electromagnetic pulse propagation distance prediction model combines the multi-dimensional characteristics of the electromagnetic pulse signal. The model parameters are trained using high-precision multi-station lightning locating data and electromagnetic pulse waveform data to obtain a high-precision electromagnetic pulse propagation distance prediction model, thereby effectively improving the accuracy and precision of lightning electromagnetic pulse locating.

[0041] (3) The lightning electromagnetic wave positioning method of the present invention is based on single-station detection. The lightning electromagnetic pulse propagation distance prediction model has good propagation distance prediction capabilities for different types of lightning electromagnetic pulse signals. The model is highly portable and applicable to a variety of computing platforms. The model has a fast calculation speed and can realize real-time detection of lightning electromagnetic pulses. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0043] Figure 1 It is a flow chart of a lightning electromagnetic wave location method based on single-station detection according to the present invention;

[0044] Figure 2 It is a schematic diagram of the signal collection and processing flow of an embodiment of a lightning electromagnetic wave location method based on single-station detection of the present invention;

[0045] Figure 3 It is a schematic diagram of the antenna structure of an embodiment of a lightning electromagnetic wave location method based on single-station detection of the present invention;

[0046] Figure 4 It is a schematic diagram of the electromagnetic field waveform of a lightning electromagnetic pulse collected by an embodiment of a lightning electromagnetic wave locating method based on single-station detection of the present invention;

[0047] Figure 5 It is a structural diagram of a lightning electromagnetic pulse propagation distance prediction model of an embodiment of a lightning electromagnetic wave locating method based on single-station detection of the present invention;

[0048] Figure 6 It is a branch network structure diagram of a multi-dimensional feature extraction network of an embodiment of a lightning electromagnetic wave location method based on single-station detection of the present invention;

[0049] Figure 7 It is a multi-feature fusion and mapping network structure diagram of an embodiment of the lightning electromagnetic wave positioning method based on single-station detection of the present invention. DETAILED DESCRIPTION

[0050] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0051] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0052] The present invention provides a lightning electromagnetic wave positioning method based on single-station detection. The method combines the electromagnetic pulse signal arrival azimuth calculation and propagation distance prediction method, wherein the electromagnetic pulse propagation distance prediction method is based on high-precision lightning positioning big data and artificial intelligence models. The method systematically solves the problems of low positioning accuracy, low detection efficiency, small monitoring range, and poor environmental adaptability of existing single-station lightning detection equipment; the method fully considers the many factors that affect the prediction of lightning electromagnetic pulse propagation distance, especially the impact of different lightning types on propagation distance prediction.

[0053] The present invention provides a lightning electromagnetic wave positioning method based on single-station detection, the positioning method comprising:

[0054] Step S10, collecting lightning electromagnetic pulse waveform data in real time through a lightning electromagnetic pulse detection device, and demodulating the received GPS or Beidou timing signal into high-precision timing information;

[0055] Step S20, based on the collected lightning electromagnetic pulse waveform data and the high-precision timing information, and using a trained lightning electromagnetic pulse propagation distance prediction model to calculate the propagation distance of the lightning electromagnetic pulse;

[0056] Step S30, calculating the wave direction of the lightning electromagnetic pulse based on the collected lightning electromagnetic pulse waveform data;

[0057] Step S40, fusing the lightning electromagnetic pulse propagation distance and wave direction to obtain the lightning electromagnetic pulse positioning result.

[0058] In order to more clearly illustrate the lightning electromagnetic wave positioning method based on single-station detection of the present invention, the following is combined with Figure 1 Each step in the embodiment of the present invention is described in detail.

[0059] The lightning electromagnetic wave location method based on single-station detection in the first embodiment of the present invention includes steps S10 to S40, each of which is described in detail as follows:

[0060] Step S10, collecting lightning electromagnetic pulse waveform data in real time through lightning electromagnetic pulse detection equipment, and demodulating the received GPS or Beidou timing signal into high-precision timing information.

[0061] In one embodiment of the present invention, the lightning electromagnetic pulse detection equipment distributed in each meteorological observation station operates in the very low frequency / low frequency (VLF / LF) frequency band, such as Figure 2 As shown, it is a schematic diagram of the signal acquisition and processing flow of an embodiment of the lightning electromagnetic wave locating method based on single-station detection of the present invention, and the 3dB frequency range of the signal acquisition channel is 3kHz to 400kHz.

[0062] The sensed analog signal is amplified and filtered, and then converted into a digital signal by an analog / digital (A / D) converter for subsequent processing. The sampling rate of the A / D converter is 1MSPS, which means that 1×10 6 The sampling precision is 16 bits.

[0063] like Figure 3 As shown, it is a schematic diagram of the antenna structure of an embodiment of the lightning electromagnetic wave locating method based on single-station detection of the present invention. The signal receiving antenna uses a flat capacitor antenna and an orthogonal magnetic loop antenna, wherein the flat capacitor antenna senses electric field signals, and the orthogonal magnetic loop antenna senses east-west magnetic field signals and north-south magnetic field signals respectively.

[0064] like Figure 4 As shown, it is a schematic diagram of the collected lightning electromagnetic pulse electromagnetic field waveform of an embodiment of the lightning electromagnetic wave locating method based on single-station detection of the present invention. The signal collection mode is level-triggered sampling, the single sampling time length is 1ms, and the pre-trigger sampling length is 120μs.

[0065] Step S20, based on the collected lightning electromagnetic pulse waveform data and the high-precision timing information, and using a trained lightning electromagnetic pulse propagation distance prediction model, calculate the propagation distance of the lightning electromagnetic pulse.

[0066] The lightning electromagnetic pulse propagation distance prediction model is trained as follows:

[0067] Step B10, obtaining a set amount of lightning electromagnetic pulse waveform data and corresponding high-precision timing information as a model training data set, and using the actual propagation distance of the lightning electromagnetic pulse corresponding to the waveform data as a label for the training data.

[0068] In recent years, a large amount of raw data of lightning electromagnetic pulses has been obtained through lightning electromagnetic pulse detection equipment distributed in various field observation stations. These detection devices form a multi-station lightning location network through networking, which can accurately calculate the location, time, intensity and other information of lightning. Furthermore, the raw electromagnetic pulse data can be marked according to the calculation results, and the marking content includes the signal propagation distance, arrival time, and azimuth information, thereby forming a lightning electromagnetic pulse waveform database.

[0069] The waveform database plays an important role in improving the positioning effect of the single-station lightning detection model. At present, there are nearly 800,000 labeled lightning electromagnetic pulse waveform data samples, including various types of lightning pulse signals, such as cloud-to-ground flashes, cloud flashes, pocket cloud flashes, etc. The sample distance range is [0, 4000km], which fully covers the range of lightning ground wave propagation.

[0070] Step B20, select any training data, and obtain the corresponding lightning electromagnetic pulse propagation distance through the constructed lightning electromagnetic pulse propagation distance prediction model.

[0071] The prediction model relies on the established lightning electromagnetic pulse waveform database for parameter training. It is necessary to read information for each sample in the database, obtain electromagnetic pulse waveform data and labeling information respectively, to facilitate subsequent model training.

[0072] In one embodiment of the present invention, a batch of data is selected from the model training data set in each iterative training, and one batch includes 1024 training data.

[0073] Step B30, calculating the loss value between the propagation distance of the lightning electromagnetic pulse and the label of the current training data through a set loss function.

[0074] During the model training process, there are usually two methods for calculating the loss value, one is to calculate the absolute error between the predicted distance and the actual distance, and the other is to calculate the relative error between the predicted distance and the actual distance. Considering that the data sample distance interval is [0, 4000km], the interval span is very large, and the use of relative error can better reflect the model's prediction ability for samples of different distances. Therefore, the present invention uses relative error to calculate the model loss value. After obtaining the loss value, the model parameters can be optimized by the neural network back propagation calculation method. The present invention uses the stochastic gradient descent optimization algorithm to update the model parameters.

[0075] Step B40, updating the parameters of the lightning electromagnetic pulse propagation distance prediction model through the optimization function, and performing iterative model training through the training data set until the set training end condition is reached, thereby obtaining a trained lightning electromagnetic pulse propagation distance prediction model.

[0076] The lightning electromagnetic pulse propagation distance prediction model is built based on a deep learning neural network, including an input layer, a multi-dimensional feature extraction network, a multi-feature fusion and mapping network, and an output layer.

[0077] like Figure 5 As shown in FIG. 1 , a lightning electromagnetic pulse propagation distance prediction model structure diagram of an embodiment of a lightning electromagnetic wave location method based on single-station detection of the present invention is shown. The input layer transmits lightning electromagnetic pulse data to the next calculation layer. The input data includes electric field data D E , north-south magnetic field data East-West magnetic field data Signal arrival time D T , the multi-dimensional feature extraction network extracts the electric field data D E , north-south magnetic field data East-West magnetic field data The features extracted are combined with the signal arrival time D by the multi-feature fusion and mapping network. T After fusion, the output layer finally outputs the predicted distance.

[0078] In one embodiment of the present invention, the electric field data D E , north-south magnetic field data and east-west magnetic field data The dimensions of are all 1×1000, and the signal arrival time D T The dimension of the high-precision timing information is 1 × 1. The signal arrival time refers to the time difference between the time when the signal arrives at the detection station and the time at midnight of the day. The time unit is seconds, that is, the signal arrival time ranges from 0 to 86400 seconds.

[0079] The multi-dimensional feature extraction network includes three branch networks with the same structure, which are used to extract the electric field signal, north-south magnetic field signal and east-west magnetic field signal of lightning electromagnetic pulse respectively.

[0080] like Figure 6 As shown, this is a branch network structure diagram of a multi-dimensional feature extraction network of an embodiment of a lightning electromagnetic wave locating method based on single-station detection of the present invention. The branch network includes a first input layer, multiple one-dimensional convolutional neural network computing layers, a batch normalization computing layer, an activation computing layer, a residual block, a maximum pooling computing layer, a global mean pooling computing layer and a first output layer.

[0081] The first input layer is used to transmit the collected lightning electromagnetic pulse data with a dimension of 1×N, where N is the sampling rate of the lightning electromagnetic pulse detection equipment and the time length of a single sampling. In one embodiment of the present invention, the multi-dimensional feature extraction branch network input layer data refers to the electric field data D E , north-south magnetic field data and east-west magnetic field data One of the output layer data corresponds to the electric field feature A E , north-south magnetic field characteristics and east-west magnetic field characteristics The dimension of the output data is 1×F, where F is the number of features output by each branch network.

[0082] One-dimensional convolutional neural network is mainly used to perceive the local features of input signals. The size of local feature sensing is determined by the size of the convolution kernel. In the embodiment of the present invention, the size of the convolution kernel is set to 1×3. Each layer of one-dimensional convolutional neural network computing unit contains a certain number of filter channels, that is, a certain number of independent convolution computing units, which are used to perceive the features of different dimensions of the signal. In the embodiment of the present invention, the number of filters is set to 32 or 16. F Represents the number of filters, N i Represents the feature length input to the one-dimensional convolutional neural network. After the one-dimensional convolutional neural network is calculated, the data dimension becomes N i ×N F .

[0083] The batch normalization calculation layer normalizes each batch of data, that is, normalizes all sample values ​​of each feature into data with a mean of 0 and a variance of 1, ensuring that the values ​​in the training data are of the same magnitude, making the model more numerically stable during training. The input data of the batch normalization calculation layer is the output data of the one-dimensional convolutional neural network calculation layer, and the output result of the batch normalization calculation layer is passed to the activation calculation layer.

[0084] The activation calculation layer performs an activation operation on the input data, introduces nonlinear factors, maps the input features to a high-dimensional nonlinear interval for interpretation, and solves problems that cannot be solved by linear models. In an embodiment of the present invention, the activation calculation layer sets data with a value less than 0 to 0, and keeps data with a value greater than or equal to 0 unchanged. This method is simple in calculation and can overcome the problem of gradient disappearance, thereby speeding up training.

[0085] In an embodiment of the present invention, the residual block R1 is composed of a set number of maximum pooling calculation layers, one-dimensional convolution calculation layers, batch normalization calculation layers and activation layers stacked in a certain order. The most important point of the residual block is that the output features of the maximum pooling calculation layer are superimposed with the output features of the residual block. This method can pass certain significant features of the upper network calculation to the deep network, so that the model can preserve good performance.

[0086] The maximum pooling calculation layer calculates the maximum value of the input data according to the set step size. This operation downsamples the input data, reduces the data dimension, and reduces the amount of model calculation.

[0087] The global mean pooling calculation layer is connected to the output layer of the multi-dimensional feature extraction branch network to perform mean calculation on the feature data of each filter channel. In the embodiment of the present invention, the data dimension of the output feature of each feature extraction branch network is 1×16, that is, 16 feature quantities are used to represent a component signal of the lightning electromagnetic pulse.

[0088] like Figure 7 As shown, it is a multi-feature fusion and mapping network structure diagram of an embodiment of the lightning electromagnetic wave positioning method based on single-station detection of the present invention. The multi-feature fusion and mapping network includes a second input layer, a data fusion layer, a residual neural network, a fully connected layer, an activation layer and a second output layer. The second input layer is used to transmit high-precision timing information and feature information of electric field signals, north-south magnetic field signals, and east-west magnetic field signals with dimensions of 1×F output by the three branch networks of the multi-dimensional feature extraction network respectively. The second output layer is used to output fused data with a dimension of 1×(3×F+1).

[0089] In an embodiment of the present invention, characteristic information of the electric field signal, the north-south magnetic field signal, the east-west magnetic field signal, and the signal arrival time are spliced ​​through a data fusion layer, and the fused data dimension is 1×(3×16+1).

[0090] The residual block R2 is composed of multiple fully connected layers and an activation layer stacked in sequence. The output features of the first-level fully connected layer of the residual block R2 are fused with the output features of the activation layer to facilitate the transfer of certain important features to the deep network.

[0091] At the end of the feature fusion and mapping network, a 1×1 fully connected layer and a linear activation layer are used to convert the feature dimension to 1×1, which is the propagation distance corresponding to the input signal predicted by the model.

[0092] Step S30, based on the collected lightning electromagnetic pulse waveform data, calculate the wave direction of the lightning electromagnetic pulse, as shown in equations (1) and (2):

[0093]

[0094]

[0095] Among them, θ is the wave direction of the lightning electromagnetic pulse, E is the maximum amplitude of the electric field signal of the lightning electromagnetic pulse, and B NS is the maximum amplitude of the north-south magnetic field signal of the lightning electromagnetic pulse, B EW It is the maximum amplitude of the east-west magnetic field signal of the lightning electromagnetic pulse.

[0096] Step S40, fusing the lightning electromagnetic pulse propagation distance and wave direction to obtain the lightning electromagnetic pulse positioning result.

[0097] The location of the detection station is known, so after obtaining the signal direction and distance information, the location of the lightning can be calculated through a simple geographic formula.

[0098] The present invention fully considers the influence of different types of lightning on the propagation distance prediction. In the training data, a large number of different types of lightning electromagnetic pulse samples are marked, and the prediction model has good learning ability for different types of data.

[0099] In practical applications, the method proposed in the present invention has good applicability. After comparing and analyzing the positioning effect of a single-station device with the positioning result of high-precision multi-station network detection, the positioning error of the single-station device is shown in Table 1:

[0100] Table 1

[0101]

[0102] As can be seen from Table 1, the lightning electromagnetic wave positioning method based on single-station detection proposed in the present invention designs a lightning electromagnetic pulse propagation distance prediction model based on a deep learning method, combines the real lightning waveform data and event information detected by high-precision multi-station networking to train the prediction model, and realizes high-precision detection of lightning events within a range of thousands of kilometers by single-station detection equipment, with a relative positioning error average value better than 5%. Compared with existing detection technologies, the method proposed in the present invention has significant improvements in positioning accuracy, detection efficiency, environmental adaptability and other aspects.

[0103] In addition, it can be seen from the data in the table that, as the distance of lightning signal transmission increases, the positioning accuracy of the lightning electromagnetic wave positioning method based on single-station detection of the present invention is higher. The main reasons are: 1. There are too many types of lightning radiation signals at short distances (50-130 km in the table), especially the cloud flash pulse sequence varies widely, and its distance estimation is a major problem currently faced by the world. The method of the present invention has reduced the positioning results of the existing technology with extremely large errors to an error of only 15.26%, which is a great progress in this field; 2. For medium distances of several hundred kilometers (150-500 km in the table), this part of the signal is mainly composed of ground lightning signals, and also contains some pocket cloud flashes (a type of cloud flash with significant characteristics). These signals have obvious characteristics and are easy to be distinguished by the model and extract information related to the propagation distance for fitting. Therefore, in this propagation range, the error of the model of the present invention is improved to 8.68%, which is also a great progress compared with the existing technology; 3. As the signal propagation distance further increases to the order of thousands of kilometers (1070-1180 km in the table), the types of lightning signals are further reduced and the signal characteristics are more obvious. On the other hand, affected by the ground-ionosphere waveguide, there is a high probability that a signal containing a mixture of ground waves and sky waves will appear in the collected single signal. There are related theoretical calculation methods to calculate this mixed signal to obtain the propagation distance. It can be simply considered that the mixed signal itself carries the propagation distance information. The model of the present invention has learned this law, which further optimizes the positioning accuracy and reduces the error to 4.91%, which is also a great improvement compared with the prior art.

[0104] Although the various steps in the above embodiment are described in the above-mentioned order, those skilled in the art can understand that in order to achieve the effect of this embodiment, different steps do not have to be executed in such an order. They can be executed simultaneously (in parallel) or in a reverse order. These simple changes are within the scope of protection of the present invention.

[0105] The lightning electromagnetic wave positioning system based on single-station detection according to the second embodiment of the present invention comprises the following modules:

[0106] A signal acquisition module is configured to collect lightning electromagnetic pulse waveform data in real time through a lightning electromagnetic pulse detection device;

[0107] A distance prediction module is configured to calculate the propagation distance based on the collected lightning electromagnetic pulse waveform data by using a trained lightning electromagnetic pulse propagation distance prediction model;

[0108] A wave orientation calculation module configured to calculate the wave orientation of the lightning electromagnetic pulse based on the collected lightning electromagnetic pulse waveform data;

[0109] The fusion positioning module is configured to fuse the propagation distance and the wave direction of the lightning electromagnetic pulse to obtain the positioning result of the lightning electromagnetic pulse.

[0110] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the system described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0111] It should be noted that the lightning electromagnetic wave location system based on single-station detection provided in the above embodiment is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the modules or steps in the embodiments of the present invention can be decomposed or combined. For example, the modules in the above embodiments can be combined into one module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present invention are only for distinguishing the modules or steps and are not regarded as improper limitations of the present invention.

[0112] An electronic device according to a third embodiment of the present invention includes:

[0113] at least one processor; and

[0114] a memory communicatively connected to at least one of the processors; wherein,

[0115] The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the above-mentioned lightning electromagnetic wave positioning method based on single-station detection.

[0116] A computer-readable storage medium according to a fourth embodiment of the present invention stores computer instructions, and the computer instructions are used to be executed by the computer to implement the above-mentioned lightning electromagnetic wave locating method based on single-station detection.

[0117] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the storage device and processing device described above can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0118] Those skilled in the art should be able to appreciate that the modules and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, computer software or a combination of the two, and the programs corresponding to the software modules and method steps can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the technical field. In order to clearly illustrate the interchangeability of electronic hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in electronic hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0119] The terms "first", "second", etc. are used to distinguish similar objects rather than to describe or indicate a particular order or sequence.

[0120] The term "comprise" or any other similar term is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus / device that includes a list of elements includes not only those elements but also other elements not expressly listed, or also includes elements inherent to such process, method, article, or apparatus / device.

[0121] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A lightning electromagnetic wave location method based on single-station detection, characterized in that: The positioning method includes: Step S10, collecting lightning electromagnetic pulse waveform data in real time through a lightning electromagnetic pulse detection device, and demodulating the received GPS or Beidou timing signal into high-precision timing information; Step S20, based on the collected lightning electromagnetic pulse waveform data and the high-precision timing information, and using a trained lightning electromagnetic pulse propagation distance prediction model to calculate the propagation distance of the lightning electromagnetic pulse; Step S30, calculating the wave direction of the lightning electromagnetic pulse based on the collected lightning electromagnetic pulse waveform data; Step S40, fusing the lightning electromagnetic pulse propagation distance and wave direction to obtain a lightning electromagnetic pulse positioning result; The lightning electromagnetic pulse propagation distance prediction model, the training method of which is: Step B10, obtaining a set amount of lightning electromagnetic pulse waveform data and corresponding high-precision timing information as a model training data set, and using the actual propagation distance of the lightning electromagnetic pulse corresponding to the waveform data as a label of the training data; Step B20, selecting any training data, and obtaining the corresponding lightning electromagnetic pulse propagation distance through the constructed lightning electromagnetic pulse propagation distance prediction model; Step B30, calculating the loss value between the propagation distance of the lightning electromagnetic pulse and the label of the current training data by using a set loss function; Step B40, updating the parameters of the lightning electromagnetic pulse propagation distance prediction model through the optimization function, and performing iterative model training through the training data set until the set training end condition is reached, thereby obtaining a trained lightning electromagnetic pulse propagation distance prediction model; The wave direction of the lightning electromagnetic pulse is calculated as follows: Among them, θ is the wave direction of the lightning electromagnetic pulse, E is the maximum amplitude of the electric field signal of the lightning electromagnetic pulse, and B NS is the maximum amplitude of the north-south magnetic field signal of the lightning electromagnetic pulse, B EW It is the maximum amplitude of the east-west magnetic field signal of the lightning electromagnetic pulse.

2. The lightning electromagnetic wave location method based on single-station detection according to claim 1 is characterized in that: The lightning electromagnetic pulse propagation distance prediction model is constructed based on a deep learning neural network, including an input layer, a multi-dimensional feature extraction network, a multi-feature fusion and mapping network, and an output layer.

3. The lightning electromagnetic wave location method based on single-station detection according to claim 2 is characterized in that: The multi-dimensional feature extraction network includes three branch networks with the same structure, which are respectively used to extract the electric field signal, the north-south magnetic field signal and the east-west magnetic field signal of the lightning electromagnetic pulse; The branch network includes a first input layer, multiple one-dimensional convolutional neural network computing layers, a batch normalization computing layer, an activation computing layer, a residual block, a maximum pooling computing layer, a global mean pooling computing layer and a first output layer.

4. The lightning electromagnetic wave location method based on single-station detection according to claim 3 is characterized in that: The first input layer is used to transmit the collected lightning electromagnetic pulse data with a dimension of 1×N, where N is the sampling rate of the lightning electromagnetic pulse detection equipment and the time length of a single sampling; The first output layer is used to output feature data with a dimension of 1×F extracted by the three branch networks respectively, where F is the number of features output by each branch network.

5. The lightning electromagnetic wave location method based on single-station detection according to claim 4 is characterized in that: The multi-feature fusion and mapping network includes a second input layer, a data fusion layer, a residual neural network, a fully connected layer, an activation layer and a second output layer.

6. The lightning electromagnetic wave location method based on single-station detection according to claim 5 is characterized in that: The second input layer is used to transmit high-precision timing information and feature information of electric field signals, north-south magnetic field signals, and east-west magnetic field signals with dimensions of 1×F output by three branch networks of the multi-dimensional feature extraction network; The second output layer is used to output fused data with a dimension of 1×(3×F+1).

7. A lightning electromagnetic wave locating system based on the lightning electromagnetic wave locating method based on single-station detection according to any one of claims 1 to 6, characterized in that: The positioning system includes the following modules: A signal acquisition module is configured to collect lightning electromagnetic pulse waveform data in real time through a lightning electromagnetic pulse detection device; The timing module is configured to receive GPS or Beidou timing signals and output high-precision timing information through the signal demodulation module and pass it to the lightning electromagnetic pulse propagation distance prediction model; A distance prediction module is configured to calculate the propagation distance of the lightning electromagnetic pulse based on the collected lightning electromagnetic pulse waveform data and the high-precision timing information through a trained lightning electromagnetic pulse propagation distance prediction model; A wave orientation calculation module, configured to calculate the wave orientation of the lightning electromagnetic pulse based on the collected lightning electromagnetic pulse waveform data; The fusion positioning module is configured to fuse the propagation distance and wave direction of the lightning electromagnetic pulse to obtain the positioning result of the lightning electromagnetic pulse.

8. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to at least one of the processors; wherein, The memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the lightning electromagnetic wave positioning method based on single-station detection as described in any one of claims 1-6.