Timing signal generation method, apparatus, device, and storage medium

Abstract

Embodiments of the present application provide a time sequence signal generation method, device and equipment, and a storage medium, and relate to the technical field of signal processing. The method inputs a time sequence condition parameter into a pre-trained Gaussian point dyeing model, obtains a covariance matrix of each two-dimensional Gaussian ellipse surface based on the time sequence condition parameter, projects an ellipse center of the two-dimensional Gaussian ellipse surface to obtain a projection transformation matrix, calculates a projection transparency of each two-dimensional Gaussian ellipse surface based on the projection transformation matrix and the covariance matrix, calculates at least one target Gaussian ellipse surface corresponding to each time sequence sampling point, sorts the target Gaussian ellipse surface according to the projection transparency, calculates a signal amplitude corresponding to the time sequence sampling point according to a sorting result, and obtains a target time sequence signal based on the signal amplitude. The corresponding Gaussian point dyeing model has a simple calculation process and requires fewer parameters for calculation, can generate a time sequence signal faster under less calculation resources, and improves the generation efficiency of the time sequence signal.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of signal processing, and particularly relates to a time series signal generation method and device, equipment and a storage medium. BACKGROUND

[0002] In recent years, artificial intelligence algorithms such as large language model algorithms are applied to many fields such as finance, signal processing and power. The training process of these model algorithms usually needs massive data support to achieve better prediction effect. In the field of time series signals, such as music, physical signals and stock trends, there is a lack of sufficient time series signals as training data.

[0003] In related technologies, a simulated time series signal is generated by using a Transformer model as training data to improve the quantity and diversity of training data. However, the Transformer model is composed of a self-attention mechanism and a fully connected network, and the calculation complexity is high when generating a signal, which requires more computing resources and is slow in generation speed. SUMMARY

[0004] The main purpose of the embodiments of the present application is to provide a time series signal generation method, device, equipment and storage medium to improve the generation efficiency of time series signals.

[0005] To achieve the above purpose, a first aspect of the embodiments of the present application provides a time series signal generation method, comprising:

[0006] obtaining a time series condition parameter and a time series sampling point number, inputting the time series condition parameter into a pre-trained Gaussian point dyeing model, the Gaussian point dyeing model comprising a plurality of two-dimensional Gaussian elliptical surfaces;

[0007] obtaining a covariance matrix of each two-dimensional Gaussian elliptical surface based on the time series condition parameter, projecting an elliptical center of the two-dimensional Gaussian elliptical surface to obtain a projection transformation matrix, and calculating a projection transparency of each two-dimensional Gaussian elliptical surface based on the projection transformation matrix and the covariance matrix;

[0008] generating a time series sampling point consistent with the time series sampling point number, calculating at least one target Gaussian elliptical surface corresponding to each time series sampling point, sorting the target Gaussian elliptical surface according to the projection transparency, calculating a signal amplitude corresponding to the time series sampling point according to the sorting result, and obtaining a target time series signal based on the signal amplitude.

[0009] In some embodiments, the obtaining of the covariance matrix of each two-dimensional Gaussian elliptical surface based on the time series condition parameter comprises:

[0010] determining a scaling vector of each of the two-dimensional Gaussian ellipse surface according to the timing condition parameter, a first scaling factor and a second scaling factor, the first scaling factor and the second scaling factor corresponding to a coordinate vector of an ellipse center of the two-dimensional Gaussian ellipse surface respectively;

[0011] obtaining the covariance matrix according to the scaling vector and a transpose vector of the scaling vector.

[0012] In some embodiments, the projecting the ellipse center of the two-dimensional Gaussian ellipse surface to obtain a projection transformation matrix comprises:

[0013] projecting based on the coordinate vector of the ellipse center to obtain a corresponding projection vector;

[0014] calculating a Jacobian matrix between the projection vector and the coordinate vector, and taking the Jacobian matrix as the projection transformation matrix.

[0015] In some embodiments, the calculating the projection transparency of each of the two-dimensional Gaussian ellipse surface based on the projection transformation matrix and the covariance matrix comprises:

[0016] obtaining a scaling transformation matrix according to the projection transformation matrix and the scaling vector;

[0017] obtaining a projection covariance matrix according to the scaling transformation matrix, the covariance matrix and a transpose matrix of the scaling transformation matrix;

[0018] obtaining a target transparency of the two-dimensional Gaussian ellipse surface, and calculating a difference vector between the projection vector and the coordinate vector;

[0019] calculating the projection transparency based on the target transparency, the difference vector and an inverse matrix of the projection covariance matrix.

[0020] In some embodiments, the calculating at least one target Gaussian ellipse surface corresponding to each of the timing sampling points comprises:

[0021] obtaining a projection interval corresponding to each of the two-dimensional Gaussian ellipse surfaces;

[0022] traversing the projection interval, and if the timing sampling point is located in the projection interval, taking the two-dimensional Gaussian ellipse surface as the target Gaussian ellipse surface corresponding to the timing sampling point.

[0023] In some embodiments, the sorting the target Gaussian ellipse surface according to the projection transparency and calculating a signal amplitude corresponding to the timing sampling point according to a sorting result comprises:

[0024] obtaining a Gaussian ellipse sequence according to the projection transparencies of the target Gaussian ellipsoids from small to large, and obtaining an ellipse number corresponding to the Gaussian ellipse sequence;

[0025] selecting a target number according to the ellipse number from small to large, sequentially selecting the target Gaussian ellipsoids smaller than the target number in the Gaussian ellipse sequence, and obtaining a first target parameter based on the corresponding projection transparencies;

[0026] obtaining a reference Gaussian ellipsoid corresponding to the target number in the Gaussian ellipse sequence, and obtaining a second target parameter according to the projection transparency and a weight factor of the reference Gaussian ellipsoid;

[0027] obtaining a target parameter corresponding to the target number according to the product of the first target parameter and the second target parameter, and accumulating the target parameter to obtain the signal amplitude.

[0028] In some embodiments, the training process of the Gaussian point dyeing model includes the following steps:

[0029] generating a plurality of reference timing signals based on a reference sampling point number and different reference timing condition parameters, the reference timing signals including sampling signal amplitudes consistent with the reference sampling point number;

[0030] generating an initial parameter of each two-dimensional Gaussian ellipsoid in an initial Gaussian point dyeing model, the initial parameter including at least an initial weight factor, an initial transparency, a first initial scaling factor, a second initial scaling factor, and an initial coordinate vector of a corresponding ellipse center;

[0031] inputting the reference timing condition parameters into the initial Gaussian point dyeing model to generate timing signals, respectively, to obtain corresponding predicted timing signals, the predicted timing signals including predicted signal amplitudes consistent with the reference sampling point number;

[0032] calculating a loss value based on the predicted signal amplitudes and the sampling signal amplitudes, adjusting the initial parameter based on the loss value until an iteration termination condition is reached, and obtaining a trained initial Gaussian point dyeing model.

[0033] To achieve the above object, a second aspect of the embodiments of the present application proposes a timing signal generation device, comprising:

[0034] a data input module for obtaining timing condition parameters and timing sampling point numbers, inputting the timing condition parameters into a pre-trained Gaussian point dyeing model, and the Gaussian point dyeing model including a plurality of two-dimensional Gaussian ellipsoids;

[0035] The parameter calculation module is configured to obtain a covariance matrix of each two-dimensional Gaussian elliptical surface based on the timing condition parameter, project an elliptical center of the two-dimensional Gaussian elliptical surface to obtain a projection transformation matrix, and calculate a projection transparency of each two-dimensional Gaussian elliptical surface based on the projection transformation matrix and the covariance matrix.

[0036] The signal generation module is configured to generate timing sampling points consistent with the number of timing sampling points, calculate at least one target Gaussian elliptical surface corresponding to each timing sampling point, sort the target Gaussian elliptical surface according to the projection transparency, calculate a signal amplitude corresponding to the timing sampling point according to a sorting result, and obtain a target timing signal based on the signal amplitude.

[0037] To achieve the above object, a third aspect of the embodiments of the present application provides an electronic device, which comprises a memory and a processor, the memory stores a computer program, and the processor implements the method of the first aspect when executing the computer program.

[0038] To achieve the above object, a fourth aspect of the embodiments of the present application provides a storage medium, which is a storage medium, the storage medium stores a computer program, and the computer program is executed by a processor to implement the method of the first aspect.

[0039] The timing signal generation method, device, equipment and storage medium provided by the embodiments of the present application obtain timing condition parameters and the number of timing sampling points, input the timing condition parameters into a pre-trained Gaussian point dyeing model, obtain a covariance matrix of each two-dimensional Gaussian elliptical surface based on the timing condition parameters, project an elliptical center of the two-dimensional Gaussian elliptical surface to obtain a projection transformation matrix, calculate a projection transparency of each two-dimensional Gaussian elliptical surface based on the projection transformation matrix and the covariance matrix, generate timing sampling points consistent with the number of timing sampling points, calculate at least one target Gaussian elliptical surface corresponding to each timing sampling point, sort the target Gaussian elliptical surface according to the projection transparency, calculate a signal amplitude corresponding to the timing sampling point according to a sorting result, and obtain a target timing signal based on the signal amplitude. The calculation process of the Gaussian point dyeing model in the embodiments of the present application is simple, and fewer parameters are required for calculation, so that the timing signal can be generated faster under less computing resources, and the generation efficiency of the timing signal is improved. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 is a flowchart of the timing signal generation method provided by the embodiments of the present application.

[0041] Figure 2 is a flowchart of obtaining a covariance matrix of each two-dimensional Gaussian elliptical surface based on timing condition parameters provided by the embodiments of the present application.

[0042] Figure 3 is a flowchart provided by an embodiment of the present application for obtaining a projection transformation matrix by projecting an ellipse center of a two-dimensional Gaussian elliptic surface.

[0043] Figure 4 is a flowchart provided by an embodiment of the present application for calculating a projection transparency of each two-dimensional Gaussian elliptic surface based on a projection transformation matrix and a covariance matrix.

[0044] Figure 5 is a flowchart provided by an embodiment of the present application for calculating at least one target Gaussian elliptic surface corresponding to each time sequence sampling point.

[0045] Figure 6 is a projection schematic diagram of a two-dimensional Gaussian elliptic surface in an embodiment of the present application.

[0046] Figure 7 is a flowchart provided by an embodiment of the present application for sorting target Gaussian elliptic surfaces according to a projection transparency, and calculating a signal amplitude corresponding to a time sequence sampling point according to a sorting result.

[0047] Figure 8 is a training process flowchart of a Gaussian point dyeing model provided by an embodiment of the present application.

[0048] Figure 9 is a training flowchart schematic diagram provided by an embodiment of the present application.

[0049] Figure 10 is a training effect schematic diagram provided by an embodiment of the present application.

[0050] Figure 11 is a structure block diagram of a time sequence signal generation device provided by another embodiment of the present application.

[0051] Figure 12 is a hardware structure schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0053] It should be noted that although the functional modules are divided in the device schematic diagram, and the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a manner different from the module division in the device or the order in the flowchart.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to be limiting of this application.

[0055] In recent years, artificial intelligence algorithms such as large language model algorithms are applied to many fields such as finance, signal processing, and power. The training process of these model algorithms usually needs massive data support to achieve better prediction effect. In the field of time series signals, such as music, physical signals, and stock trends, there is a lack of enough time series signals as training data.

[0056] In the related art, a simulated time series signal is generated by using a Transformer model as training data to improve the quantity and diversity of the training data. However, the Transformer model is composed of a self-attention mechanism and a fully connected network, and when generating a signal, the computational complexity is high, more computing resources are needed, and the generation speed is slow.

[0057] Therefore, the embodiments of the present application provide a time series signal generation method, device, equipment and storage medium, wherein the calculation process of the Gaussian point dyeing model is simple, and fewer parameters are needed for calculation, so that the time series signal can be generated faster under less computing resources, and the generation efficiency of the time series signal is improved.

[0058] The embodiments of the present application provide a time series signal generation method, device, equipment and storage medium, which are specifically described by the following embodiments. First, the time series signal generation method in the embodiments of the present application is described.

[0059] The time series signal generation method provided by the embodiments of the present application relates to the technical field of signal processing. The time series signal generation method provided by the embodiments of the present application can be applied in a terminal, can be applied in a server, and can also be a computer program running in the terminal or the server. For example, the computer program can be a native program or a software module in an operating system; can be a native application program (APP), that is, a program that needs to be installed in an operating system to run, such as a client supporting time series signal generation, or a small program that only needs to be downloaded into a browser environment to run; or a small program that can be embedded into any APP. In summary, the above computer program can be any form of application program, module or plug-in. The terminal communicates with the server through a network. The time series signal generation method can be executed by the terminal or the server, or cooperatively executed by the terminal and the server.

[0060] In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart watch, or the like. In addition, the terminal can also be a smart vehicle-mounted device. The smart vehicle-mounted device applies the timing signal generation method of the embodiments to provide related services and improve the driving experience. The server can be a standalone server, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and basic cloud computing services such as big data and artificial intelligence platforms; or a service node in a blockchain system, the service nodes in the blockchain system form a peer-to-peer (P2P) network, and the P2P protocol is an application layer protocol running on the transmission control protocol (TCP) protocol. The terminal and the server can be connected through a communication connection mode such as Bluetooth, universal serial bus (USB), or a network, and the embodiments are not limited herein.

[0061] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment, in which tasks are performed by remote processing devices connected by a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0062] The timing signal generation method in the embodiments of the present application is described below.

[0063] Figure 1 is an optional flowchart of the timing signal generation method provided by the embodiments of the present application, Figure 1 The method in can include but is not limited to steps 110 to 130. It can be understood that the embodiments of the present application do not make specific limitations on the order of steps 110 to 130 in Figure 1 Steps 110 to 130 in the embodiments of the present application are not limited in the order, and the order of steps can be adjusted or some steps can be reduced, added, or the like according to actual needs.

[0064] Step 110: Obtain the timing condition parameter and the timing sampling point number, and input the timing condition parameter into the pre-trained Gaussian point dyeing model.

[0065] In an embodiment, the timing condition parameter is a characteristic parameter used to characterize the required timing signal, and the characteristic parameters of the timing signal are different in different fields. For example, in the music field, the timing condition parameter can be the parameter corresponding to the music style, and the timing music signal can be generated according to the specified music style. In the financial field, the timing condition parameter can be a parameter describing the rise and fall, and a timing signal representing the stock trend can be generated according to the rise and fall description; in the field of physical science, the timing condition parameter can be the characteristic of celestial bodies, and the timing gravitational wave signal can be generated according to the characteristic of celestial bodies.

[0066] The embodiment of the present application selects the field of astrophysics and generates a binary black hole merging timing signal as a specific application scenario for introduction. In an embodiment, taking a binary black hole (BBH) as an example, it is assumed that the mass ratio of the two black holes is 10 and the tilt angle between the two black holes is θ. As θ changes, the signal waveform of the corresponding gravitational wave between the two black holes is also different. At this time, the tilt angle θ is taken as the timing condition parameter.

[0067] In an embodiment, the timing signal is a discrete signal, which is composed of sampling points and signal amplitudes corresponding to the sampling points. Therefore, the timing sampling point number is related to the total length of the timing signal. For example, if the sampling rate is 200 Hz and the timing sampling point number is 20,000, the corresponding signal length is 100 s.

[0068] In an embodiment, the Gaussian point dyeing model is a set composed of gsn two-dimensional Gaussian elliptical surfaces. The more the number of two-dimensional Gaussian elliptical surfaces, the more accurate the Gaussian point dyeing process, and the more accurate the generated target timing signal. For example, the number of two-dimensional Gaussian elliptical surfaces is gsn=10000. It can be understood that the parameters of the Gaussian point dyeing model in the embodiment of the present application are the parameters corresponding to each two-dimensional Gaussian elliptical surface, and the training process is completed in advance to obtain the trained Gaussian point dyeing model.

[0069] In the above embodiment, after obtaining the timing condition parameter and the timing sampling point number, the timing condition parameter can be input into the pre-trained Gaussian point dyeing model for prediction, and then the final target timing signal is generated.

[0070] Step 120: Obtain the covariance matrix of each two-dimensional Gaussian elliptical surface based on the timing condition parameter, project the elliptical center of the two-dimensional Gaussian elliptical surface to obtain a projection transformation matrix, and calculate the projection transparency of each two-dimensional Gaussian elliptical surface based on the projection transformation matrix and the covariance matrix.

[0071] In an embodiment, each two-dimensional Gaussian elliptical surface includes a coordinate vector t of an elliptical center, a first scaling factor s1 and a second scaling factor s2 corresponding to two values of the coordinate vector, a weight factor c indicating a proportion of importance of the two-dimensional Gaussian elliptical surface in the Gaussian point dyeing model, and a target transparency a used in the calculation process of the two-dimensional Gaussian elliptical surface.

[0072] wherein the coordinate vector t is expressed as: Since the coordinate vector is a two-dimensional coordinate, t0 and t1 represent two values corresponding to different coordinate axes, respectively.

[0073] Therefore, the parameters contained in each two-dimensional Gaussian elliptical surface can be represented as {a, c, s1, s2, t0, t1}. If the Gaussian point dyeing model is represented as GS, the Gaussian point dyeing model GS contains gsn two-dimensional Gaussian elliptical surfaces, and therefore the amount of parameters contained in the Gaussian point dyeing model is: 6*gsn=60000. According to the definition of two-dimensional Gaussian distribution, each two-dimensional Gaussian elliptical surface can also be characterized by a covariance matrix ∑, and the two-dimensional Gaussian elliptical surface is expressed as:

[0074]

[0075] wherein ∑ -1 represents the inverse matrix of the covariance matrix, h represents an input parameter, and T represents transposition.

[0076] At this time, the Gaussian point dyeing model is expressed as:

[0077] GS={G(∑ i ,α i ,c i ,s1 i ,s2 i ,t0 i ,t1 i )|i∈[1,gsn]}

[0078] The parameters contained in the Transformer algorithm in the related art are usually in the order of millions. It can be seen that the Gaussian point dyeing model adopted in the embodiments of the present application can greatly reduce the amount of parameters, and therefore the computing resources required for calculation are also greatly reduced, and the training speed and inference speed of the model are also improved.

[0079] In an embodiment, the time sequence condition parameter is a tilt angle θ, and the time sequence condition parameter is substituted into each two-dimensional Gaussian elliptical surface for calculation. First, the covariance matrix of each two-dimensional Gaussian elliptical surface is calculated. Referring to Figure 2 , Figure 2 is a flowchart provided by the embodiments of the present application for obtaining the covariance matrix of each two-dimensional Gaussian elliptical surface based on the time sequence condition parameter, and specifically includes steps 210 to 220:

[0080] Step 210: determining a scaling vector of each two-dimensional Gaussian ellipse surface according to the timing condition parameter, the first scaling factor and the second scaling factor.

[0081] In an embodiment, the input timing condition parameter is embedded by the scaling vector, and the calculation process of the specific scaling vector W is as follows:

[0082] W = RS

[0083]

[0084] Step 220: obtaining a covariance matrix according to the scaling vector and the transpose vector of the scaling vector.

[0085] In an embodiment, the covariance matrix is expressed as:

[0086] ∑ = RSS T R T = WW T

[0087] wherein W T represents the transpose vector of the scaling vector.

[0088] In an embodiment, it is also required to project the two-dimensional Gaussian ellipse surface to a one-dimensional timing sampling space, and therefore it is required to calculate a projection transformation matrix. Referring to Figure 3 , Figure 3 is a flowchart provided by the embodiment of the application for projecting the ellipse center of the two-dimensional Gaussian ellipse surface to obtain a projection transformation matrix, and specifically includes steps 310 to 320:

[0089] Step 310: projecting based on the coordinate vector of the ellipse center to obtain a corresponding projection vector.

[0090] In an embodiment, the projection function is The projection vector x obtained by projecting the coordinate vector of the ellipse center based on the projection function is expressed as:

[0091]

[0092] wherein x represents the projection vector, x0 and x1 represent two values of the projection vector respectively representing different coordinate axes.

[0093] Step 320: calculating the Jacobian matrix between the projection vector and the coordinate vector, and taking the Jacobian matrix as the projection transformation matrix.

[0094] wherein the Jacobian matrix is a matrix arranged in a certain way by the first-order partial derivative of a function, and the projection transformation matrix J is expressed as:

[0095]

[0096] After the covariance matrix and the projection transformation matrix are calculated, the distribution of the two-dimensional Gaussian ellipse surface projected into the one-dimensional time sequence space can be calculated, and based on the distribution, the projection transparency of the target transparency after projection can be calculated. In an embodiment, referring to Figure 4 , Figure 4 is a flowchart for calculating the projection transparency of each two-dimensional Gaussian ellipse surface based on the projection transformation matrix and the covariance matrix, and specifically includes steps 410 to 440:

[0097] Step 410: Obtain a scaling transformation matrix according to the projection transformation matrix and the scaling vector.

[0098] In an embodiment, the scaling transformation matrix F is expressed as: F = JW.

[0099] Step 420: Obtain a projection covariance matrix according to the scaling transformation matrix, the covariance matrix, and the transpose matrix of the scaling transformation matrix.

[0100] Wherein, the projection covariance matrix ∑' is expressed as:

[0101] ∑' = F * ∑ * F T = JW∑W T J T

[0102] Step 430: Obtain the target transparency of the two-dimensional Gaussian ellipse surface, and calculate the difference vector of the projection vector and the coordinate vector.

[0103] Wherein, the difference vector is expressed as x-t.

[0104] Step 440: Calculate the projection transparency based on the target transparency, the difference vector, and the inverse matrix of the projection covariance matrix.

[0105] Wherein, the projection transparency α' is expressed as:

[0106]

[0107] Next, through the Gaussian point dyeing process, the signal amplitude corresponding to each sampling point of the target time sequence signal is calculated.

[0108] Step 130: Generate time sequence sampling points consistent with the number of time sequence sampling points, calculate at least one target Gaussian ellipse surface corresponding to each time sequence sampling point, sort the target Gaussian ellipse surface according to the projection transparency, calculate the signal amplitude corresponding to the time sequence sampling point according to the sorting result, and obtain the target time sequence signal based on the signal amplitude.

[0109] In an embodiment, the time sampling points are first generated. For example, the time sampling points are 20000, and 20000 time sampling points are set in the one-dimensional time sampling space at equal intervals.

[0110] The signal amplitude of the time sampling point j is represented as:

[0111] Since the coordinate vectors of the ellipse centers of each two-dimensional Gaussian ellipse are different, the coverage ranges of the time sampling points covered by the projected two-dimensional Gaussian ellipses are different. At this time, the coverage relationship between the time sampling points and the two-dimensional Gaussian ellipses needs to be considered. Refer to Figure 5 , Figure 5 is a flowchart for calculating at least one target Gaussian ellipse corresponding to each time sampling point provided by the embodiment of the present application, and specifically includes steps 510 to 520:

[0112] Step 510: Obtain the projection interval corresponding to each two-dimensional Gaussian ellipse.

[0113] In an embodiment, refer to Figure 6 , Figure 6 is a projection diagram of a two-dimensional Gaussian ellipse in the embodiment of the present application. The projections of the two different two-dimensional Gaussian ellipses on the horizontal coordinate axis (equivalent to the one-dimensional time sampling space) correspond to different coverage ranges, that is, different projection intervals. Specifically, according to the ellipse equation of the two-dimensional Gaussian ellipse, the interval formed by the vertical projections of all points on the ellipse on the horizontal coordinate axis is selected as the projection interval.

[0114] Step 520: Traverse the projection interval. If the time sampling point is located in the projection interval, the two-dimensional Gaussian ellipse is taken as the target Gaussian ellipse corresponding to the time sampling point.

[0115] In an embodiment, refer to Figure 6 The two-dimensional Gaussian ellipse 1 corresponds to the projection interval 1, and the two-dimensional Gaussian ellipse 2 corresponds to the projection interval 2. If there are time sampling points j=n and j=m, it can be seen that the time sampling point is located in the projection interval 1, and the time sampling point is located in the projection interval 1 and the projection interval 2 at the same time. Therefore, for the time sampling point n, it has one target Gaussian ellipse, which is the two-dimensional Gaussian ellipse 1, and for the time sampling point m, it has two target Gaussian ellipses, which are the two-dimensional Gaussian ellipse 1 and the two-dimensional Gaussian ellipse 2.

[0116] According to the above process, at least one target Gaussian ellipse corresponding to each time sampling point is obtained. Then, the signal amplitude corresponding to each time sampling point is obtained based on the target Gaussian ellipse. In an embodiment, refer to Figure 7 ,Figure 7 is a flowchart provided by an embodiment of the present application for sorting target Gaussian elliptical surfaces according to projection transparency, and calculating signal amplitudes corresponding to time sequence sampling points according to the sorting result, and specifically includes steps 710 to 740:

[0117] Step 710: Obtain a Gaussian elliptical surface sequence in order of projection transparency from small to large of each target Gaussian elliptical surface, and obtain the number of elliptical surfaces corresponding to the Gaussian elliptical surface sequence.

[0118] In an embodiment, the time sequence sampling point n is The set of target Gaussian elliptical surfaces corresponding to the time sequence sampling point n is represented as The number of elliptical surfaces of the target Gaussian elliptical surface contained in the set is For example, in the set Figure 6 The number of elliptical surfaces of the time sequence sampling point n is: The number of elliptical surfaces of the time sequence sampling point m is:

[0119] Since each two-dimensional Gaussian elliptical surface has a corresponding projection transparency, the projection transparency represents the reflection intensity after the light passes through the corresponding target Gaussian elliptical surface, thereby affecting the signal amplitude after the target Gaussian elliptical surface is projected. Therefore, for each time sequence sampling point, the present embodiment arranges all target Gaussian elliptical surfaces corresponding to the time sequence sampling point in order of projection transparency from small to large, obtains a Gaussian elliptical surface sequence, and obtains the number of elliptical surfaces corresponding to the Gaussian elliptical surface sequence.

[0120] It can be understood that the Gaussian elliptical surface sequence and the set contain the same target Gaussian elliptical surface, and the number of elliptical surfaces of the two is consistent.

[0121] In an embodiment, the set includes target Gaussian elliptical surface r1 (projection transparency is 0.6), target Gaussian elliptical surface r2 (projection transparency is 0.3), target Gaussian elliptical surface r3 (projection transparency is 0.5), target Gaussian elliptical surface r4 (projection transparency is 0.2), and target Gaussian elliptical surface r5 (projection transparency is 0.8). At this time, in order of projection transparency from small to large, the obtained Gaussian elliptical surface sequence is represented as: {target Gaussian elliptical surface r4, target Gaussian elliptical surface r2, target Gaussian elliptical surface r3, target Gaussian elliptical surface r1, target Gaussian elliptical surface r5}, and the corresponding number of elliptical surfaces is 5.

[0122] Step 720: Select a target number in order of small to large based on the number of elliptical surfaces, and sequentially select target Gaussian elliptical surfaces smaller than the target number in the Gaussian elliptical surface sequence, and obtain a first target parameter based on the corresponding projection transparency multiplication.

[0123] In an embodiment, the target number k is selected based on the number of ellipses The target number k is selected in ascending order, and the value of the target number k is: At this time, the target Gaussian ellipsoid smaller than the target number is selected in the sequence of Gaussian ellipsoid. For example, k = 1, 0 target Gaussian ellipsoid is selected, k = 2, 1 target Gaussian ellipsoid is selected, Then, the target Gaussian ellipsoid is selected .

[0124] After selecting the target Gaussian ellipsoid smaller than the target number, the first target parameter A is obtained based on the corresponding projection transparency multiplication, which is represented as:

[0125]

[0126] Wherein, α m ′ represents the projection transparency corresponding to the mth target Gaussian ellipsoid. In the embodiment of the application, the (1-α m ′) corresponding to the k-1 target Gaussian ellipsoid is multiplied, that is, the first target parameter is obtained.

[0127] The sequence of Gaussian ellipsoid is represented as: {target Gaussian ellipsoid r4, target Gaussian ellipsoid r2, target Gaussian ellipsoid r3, target Gaussian ellipsoid r1, target Gaussian ellipsoid r5} for example. If k = 1, 0 target Gaussian ellipsoid is selected, and the first target parameter is 1 at this time. When k = 2, 1 target Gaussian ellipsoid is selected, that is, target Gaussian ellipsoid r4 is selected, and the first target parameter calculated based on its projection transparency is 1. When k = 3, 2 target Gaussian ellipsoids are selected, that is, target Gaussian ellipsoid r4 and target Gaussian ellipsoid r2 are selected, when m = 1, (1-α1′) is calculated based on the projection transparency of target Gaussian ellipsoid r4, when m = 2, (1-α2′) is calculated based on the projection transparency of target Gaussian ellipsoid r2, and the first target parameter is obtained by multiplying (1-α1′) and (1-α2′). In this way, the first target parameter corresponding to each target number k is calculated.

[0128] Step 730: obtaining the target Gaussian ellipsoid corresponding to the target number in the sequence of Gaussian ellipsoid as the reference Gaussian ellipsoid, and obtaining the second target parameter according to the projection transparency and the weight factor of the reference Gaussian ellipsoid.

[0129] For example, take {target Gaussian ellipsoid r4, target Gaussian ellipsoid r2, target Gaussian ellipsoid r3, target Gaussian ellipsoid r1, target Gaussian ellipsoid r5} as an example of a Gaussian ellipsoid sequence. When k = 1, the reference Gaussian ellipsoid is the target Gaussian ellipsoid r4, when k = 2, the reference Gaussian ellipsoid is the target Gaussian ellipsoid r2, and when k = 5, the reference Gaussian ellipsoid is the target Gaussian ellipsoid r5. At this time, according to the projection transparency α of the reference Gaussian ellipsoid k and the weight factor c k to obtain the second target parameter B.

[0130] Step 740: obtaining the target parameter corresponding to the target number according to the product of the first target parameter and the second target parameter, and accumulating the target parameter to obtain the signal amplitude.

[0131] When the target number is k, the target parameter is represented as:

[0132]

[0133] At this time, the signal amplitude is obtained by accumulating the target parameter is represented as:

[0134]

[0135] According to the above process, the signal amplitude corresponding to each time sequence sampling point is calculated in turn, that is, the target time sequence signal x is obtained based on the signal amplitude θ ,

[0136] The training process of the Gaussian point dyeing model in the embodiments of the present application is described below.

[0137] In an embodiment, referring to Figure 8 , Figure 8 is the training process flowchart of the Gaussian point dyeing model provided by the embodiments of the present application, and specifically includes steps 810 to 840:

[0138] Step 810: generating a plurality of reference time sequence signals based on different reference time sequence condition parameters and reference sampling point numbers.

[0139] For example, take a binary black hole (BBH) as an example, assuming that the mass ratio of the two black holes is fixed at 10, and the inclination angle between the two black holes is changed, and the inclination angle is taken as the reference time sequence condition parameter. If the reference sampling point number is 20000, then each reference time sequence signal includes 20000 sampling points. Based on the general theory of relativity, the BBH signal waveform corresponding to each θ can be obtained, which is taken as the reference time sequence signal yθ , is represented as:

[0140]

[0141] In an embodiment, the number of signals in the training set is defined as num=50000, and a training set D containing num reference time series signals can be generated by randomly sampling num different values of θ in a uniform distribution U∈(0,π), each sample in the training set D contains a value of θ and its corresponding reference time series signal y θ , is represented as:

[0142] D={(θ i ,y i θ )|θ∈(0,π),i∈[1,num]}

[0143] Step 820: generating initial parameters of each two-dimensional Gaussian ellipse surface in the initial Gaussian point dyeing model.

[0144] In an embodiment, the initial Gaussian point dyeing model is consistent with the Gaussian point dyeing model structure, and the only difference is the parameters. The initial Gaussian point dyeing model contains initial parameters, which include: initial transparency, first initial scaling factor, second initial scaling factor, and initial coordinate vector of the corresponding ellipse center. At the beginning of training, the initial parameters are initialized, and the initial value of the initial coordinate vector is taken as an example for the i-th two-dimensional Gaussian ellipse surface. In a uniform distribution U∈[1,length], length is the signal length of the reference time series signal. The initial weight factor c i ′, the initial transparency a i ′, the first initial scaling factor The second initial scaling factor are randomly sampled in a uniform distribution U∈(0,1]. The initial parameters also include the initial covariance matrix Σ i ′ of the corresponding two-dimensional Gaussian ellipse surface. In general, the Gaussian point dyeing model of the embodiment has fewer parameters, so the convergence speed is fast and the generation speed is fast.

[0145] Step 830: inputting the reference time series condition parameters into the initial Gaussian point dyeing model respectively to generate time series signals to obtain the corresponding predicted time series signals.

[0146] In an embodiment, according to the above method of generating target time series signals, the initial Gaussian point dyeing model is used to generate the corresponding predicted time series signal y θ ′ based on the reference time series condition parameters, that is, the tilt angle. The predicted time series signal includes a predicted signal amplitude that is consistent with the number of reference sampling points, that is:

[0147]

[0148] Step 840: Calculate the loss value based on the predicted signal amplitude and the sample signal amplitude, adjust the initial parameters based on the loss value, until the iteration termination condition is reached, and obtain the trained initial Gaussian point dyeing model.

[0149] In an embodiment, the loss function can be mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), etc. The embodiment of the present application adopts mean square error (MSE) to calculate the loss value, and the loss value MSE θ is expressed as:

[0150]

[0151] wherein the loss value MSE θ is smaller, the more accurate the generated predicted time series signal is. The loss value MSE θ is smaller, the more accurate the generated predicted time series signal is. The loss value MSE is smaller, the more accurate the generated predicted time series signal is. The loss value MSE

[0152] In an embodiment, the iteration termination condition is the model convergence condition, which is specifically for all samples (θ i ,y i θ If all of them satisfy the following conditions, it is considered that the model has reached convergence:

[0153] MSE i θ <0.1, i ∈ [1, num]

[0154] Perform multiple iterations, update and optimize the initial parameters by gradient descent method, specifically, the embodiment of the present application adopts adaptive motion estimation algorithm (Adam) to update the parameters. Through repeated iteration of the above training steps, the model is trained until the Gaussian point dyeing model reaches the model convergence condition, then the training process of the Gaussian point dyeing model is completed, the model parameters are saved, and the target transparency, the first scaling factor, the second scaling factor, the coordinate vector of the corresponding ellipse center and the covariance matrix of the two-dimensional Gaussian ellipse surface are obtained.

[0155] The above training process can refer to Figure 9 , Figure 9is a training flowchart provided by an embodiment of the present application. It includes defining the number of reference sampling points to be generated, generating a training set, then defining an initial Gaussian point dyeing model and initializing the initial parameters of the model. In the process of predicting and generating a predicted time series signal, a two-dimensional Gaussian elliptical surface is projected onto a one-dimensional time series sampling space, and then the predicted amplitude of each sampling point in the predicted time series signal is calculated through Gaussian point dyeing. The loss value of the loss function is calculated based on the predicted amplitude, the model parameters corresponding to the initial Gaussian point dyeing model are optimized, and it is judged whether the model converges. If it does not converge, the prediction generation process continues. If it has converged, the model parameters are saved, and a trained Gaussian point dyeing model is obtained, thereby realizing a high-speed and accurate time series signal generation process.

[0156] In an embodiment, with reference to Figure 10 , Figure 10 is a training effect diagram provided by an embodiment of the present application. Figure 10 The predicted time series signal and the reference time series signal are shown in the figure. As can be seen from the figure, the two are relatively close, indicating that the prediction generation effect is good.

[0157] The technical scheme provided by the embodiment of the present application, by acquiring the time series condition parameters and the number of time series sampling points, inputting the time series condition parameters into the pre-trained Gaussian point dyeing model, obtaining the covariance matrix of each two-dimensional Gaussian elliptical surface based on the time series condition parameters, projecting the elliptical center of the two-dimensional Gaussian elliptical surface to obtain a projection transformation matrix, calculating the projection transparency of each two-dimensional Gaussian elliptical surface based on the projection transformation matrix and the covariance matrix, generating time series sampling points consistent with the number of time series sampling points, calculating at least one target Gaussian elliptical surface corresponding to each time series sampling point, sorting the target Gaussian elliptical surface according to the projection transparency, calculating the signal amplitude corresponding to the time series sampling point according to the sorting result, and obtaining the target time series signal based on the signal amplitude. In the embodiment of the present application, the parameter quantity and the calculation quantity of the Gaussian point dyeing model are less, and the calculation process is simple, so the target time series signal can be generated faster under less calculation resources, and the generation efficiency of the time series signal is improved.

[0158] The embodiment of the present application also provides a time series signal generation device, which can realize the above-mentioned time series signal generation method, with reference to Figure 11 , the device includes:

[0159] The data input module 1110 is configured to acquire time series condition parameters and the number of time series sampling points, and input the time series condition parameters into a pre-trained Gaussian point dyeing model. The Gaussian point dyeing model includes a plurality of two-dimensional Gaussian elliptical surfaces.

[0160] The parameter calculation module 1120 is configured to obtain a covariance matrix of each two-dimensional Gaussian elliptic surface based on the timing condition parameter, project an elliptic center of the two-dimensional Gaussian elliptic surface to obtain a projection transformation matrix, and calculate a projection transparency of each two-dimensional Gaussian elliptic surface based on the projection transformation matrix and the covariance matrix.

[0161] The signal generation module 1130 is configured to generate timing sampling points consistent with the number of timing sampling points, calculate at least one target Gaussian elliptic surface corresponding to each timing sampling point, sort the target Gaussian elliptic surface according to the projection transparency, calculate a signal amplitude corresponding to the timing sampling point according to the sorting result, and obtain a target timing signal based on the signal amplitude.

[0162] The specific implementation of the timing signal generation apparatus of the embodiment is basically the same as the specific implementation of the timing signal generation method, and will not be repeated here.

[0163] The embodiment of the present application further provides an electronic device, which comprises:

[0164] at least one memory;

[0165] at least one processor;

[0166] at least one program;

[0167] The program is stored in the memory, and the processor executes the at least one program to implement the timing signal generation method provided in the embodiment of the present application. The electronic device can be any intelligent terminal including a mobile phone, a tablet computer, a personal digital assistant (PDA), a vehicle-mounted computer, etc.

[0168] Please refer to Figure 12 , Figure 12 The hardware structure of the electronic device of another embodiment is illustrated, and the electronic device comprises:

[0169] The processor 1201 can be implemented in a general-purpose central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute related programs to implement the technical solutions provided in the embodiments of the present application.

[0170] The memory 1202 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 1202 can store an operating system and other application programs. When the technical solutions provided by the embodiments of the present specification are implemented by software or firmware, the related program codes are stored in the memory 1202 and are called and executed by the processor 1201 to implement the timing signal generation method of the embodiments of the present application.

[0171] The input / output interface 1203 is configured to realize information input and output.

[0172] The communication interface 1204 is configured to realize the communication interaction between the device and other devices. The communication can be realized by a wired manner (for example, a USB, a network cable, etc.) or a wireless manner (for example, a mobile network, WIFI, Bluetooth, etc.).

[0173] The bus 1205 is configured to transmit information between various components (for example, the processor 1201, the memory 1202, the input / output interface 1203, and the communication interface 1204) of the device.

[0174] The processor 1201, the memory 1202, the input / output interface 1203, and the communication interface 1204 are connected to each other through the bus 1205 to realize the communication connection between the device.

[0175] The embodiments of the present application also provide a storage medium. The storage medium is a storage medium, and the storage medium stores a computer program. The computer program is executed by the processor to implement the timing signal generation method.

[0176] The memory is a non-transitory storage medium, which can be used to store non-transitory software programs and non-transitory computer executable programs. In addition, the memory can include a high-speed random access memory and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory remotely arranged relative to the processor. These remote memories can be connected to the processor through a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0177] The timing signal generation method, device, equipment and storage medium provided in the embodiments of the present application, by acquiring timing condition parameters and a timing sampling point quantity, inputting the timing condition parameters into a pre-trained Gaussian point dyeing model, obtaining a covariance matrix of each two-dimensional Gaussian elliptical surface based on the timing condition parameters, projecting an elliptical center of the two-dimensional Gaussian elliptical surface to obtain a projection transformation matrix, calculating a projection transparency of each two-dimensional Gaussian elliptical surface based on the projection transformation matrix and the covariance matrix, generating timing sampling points consistent with the timing sampling point quantity, calculating at least one target Gaussian elliptical surface corresponding to each timing sampling point, sorting the target Gaussian elliptical surface according to the projection transparency, calculating a signal amplitude corresponding to the timing sampling point according to the sorting result, and obtaining a target timing signal based on the signal amplitude. In the embodiments of the present application, the parameter quantity and the calculation quantity of the Gaussian point dyeing model are less, and the calculation process is simple, so the target timing signal can be generated faster under less calculation resources, and the generation efficiency of the timing signal is improved.

[0178] The embodiments described in the embodiments of the present application are used to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0179] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and can include more or fewer steps than the figures shown, or combine certain steps or different steps.

[0180] The device embodiments described above are only schematic, and the units illustrated as separate components can or can not be physically separated, that is, can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0181] Those skilled in the art can understand that all or some steps in the above disclosed method, the functions of the modules / units in the system and the equipment can be implemented as software, firmware, hardware and their appropriate combinations.

[0182] The terms "first", "second", "third", "fourth", and the like in the description and in the claims of this application, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the use of the terms so termed is interchangeable under appropriate circumstances such that the embodiments of the application described herein are, for example, capable of orderly or chronological mundane operation, reverse order operation, based on circuitry availability, based on stated preference or the like, and that "default" or other orderings are thus permissible. Further, the terms "comprise", "comprising", "include", "including", and the like, are specifically intended to be open-ended. That is, references to individual steps and the like do not suhstantially exclude the presence of two or more of a recited step or its integral sub-steps or additional steps whether or not readily ascertainable from the description or the like. Further, the words "a" or "an", as used herein in the disclosure and elsewhere, are used indiscriminately and are to be interpreted in the same way, i.e. as meaning "one or more".

[0183] It should be understood that, in this application, "at least one" means one or more, "multiple" means two or more. "And / or", used to describe the relationship between associated objects, means that there can be three relationships, for example, "A and / or B" can mean: only A, only B, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects. "At least one of the following" or the like means any combination of these items, including single or multiple combinations. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0184] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the above units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be omitted or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed objects can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0185] The units described above as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e. they can be located in one place or distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0186] In addition, each function unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software function unit.

[0187] If the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, essentially or in part, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes multiple instructions used to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in the embodiments of the present application. The foregoing storage medium includes various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0188] The preferred embodiments of the embodiments of the present application are described above with reference to the accompanying drawings, and the scope of the rights of the embodiments of the present application is not limited thereto. Any modification, equivalent replacement and improvement made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the rights of the embodiments of the present application.

Claims

1. A method for generating a timing signal, characterized in that: include: Obtaining a timing condition parameter and the number of timing sampling points, and inputting the timing condition parameter into a pre-trained Gaussian dot-dyeing model, wherein the Gaussian dot-dyeing model includes a plurality of two-dimensional Gaussian ellipses; Determine the scaling vector of each of the two-dimensional Gaussian ellipsoids according to the timing condition parameters, a first scaling factor and a second scaling factor, wherein the first scaling factor and the second scaling factor respectively correspond to the coordinate vector of the ellipse center of the two-dimensional Gaussian ellipsoid; obtain a covariance matrix according to the scaling vector and the transposed vector of the scaling vector, and perform projection based on the coordinate vector of the ellipse center to obtain a corresponding projection vector; calculate the Jacobian matrix between the projection vector and the coordinate vector, use the Jacobian matrix as a projection transformation matrix, obtain a scaling transformation matrix according to the projection transformation matrix and the scaling vector, obtain a projection covariance matrix according to the scaling transformation matrix, the covariance matrix and the transposed matrix of the scaling transformation matrix, obtain the target transparency of the two-dimensional Gaussian ellipsoid, and calculate the difference vector between the projection vector and the coordinate vector, and obtain the projection transparency based on the target transparency, the difference vector and the inverse matrix of the projection covariance matrix; Generate time series sampling points consistent with the number of the time series sampling points, calculate at least one target Gaussian ellipse corresponding to each of the time series sampling points, sort the target Gaussian ellipse according to the projection transparency, calculate the signal amplitude corresponding to the time series sampling point according to the sorting result, and obtain the target time series signal based on the signal amplitude.

2. The timing signal generation method according to claim 1, wherein: The calculating of at least one target Gaussian ellipse corresponding to each of the time series sampling points includes: Get the projection interval corresponding to each two-dimensional Gaussian ellipse; The projection interval is traversed, and if the time series sampling point is located in the projection interval, the two-dimensional Gaussian ellipse is used as the target Gaussian ellipse corresponding to the time series sampling point.

3. The timing signal generation method according to claim 1, wherein: The step of sorting the target Gaussian ellipsoids according to the projection transparency and calculating the signal amplitudes corresponding to the time series sampling points according to the sorting results includes: Obtaining a Gaussian ellipse sequence according to the order of the projection transparency corresponding to each target Gaussian ellipse from small to large, and obtaining the number of ellipses corresponding to the Gaussian ellipse sequence; Based on the number of ellipses, a target number is selected in ascending order, target Gaussian ellipsoids smaller than the target number are selected in sequence from the Gaussian ellipsoid sequence, and a first target parameter is obtained by multiplying the projection transparency corresponding to the selected target Gaussian ellipsoids; Obtaining the target Gaussian ellipsoid with the same sequence number as the target number in the Gaussian ellipsoid sequence as a reference Gaussian ellipsoid, and obtaining a second target parameter according to the projection transparency and weight factor of the reference Gaussian ellipsoid; A target parameter corresponding to the target quantity is obtained according to the product of the first target parameter and the second target parameter, and the signal amplitude is obtained by accumulating the target parameters.

4. The timing signal generation method according to any one of claims 1 to 3, characterized in that: The training process of the Gaussian dot dyeing model includes the following steps: Based on the number of reference sampling points and different reference timing condition parameters, a plurality of reference timing signals are generated, wherein the reference timing signals include sampling signal amplitudes consistent with the number of reference sampling points; Generate initial parameters of each of the two-dimensional Gaussian ellipse surfaces in the initial Gaussian dot-dying model, the initial parameters including at least: an initial weight factor, an initial transparency, a first initial scaling factor, a second initial scaling factor, and an initial coordinate vector of the corresponding ellipse center; Inputting the reference timing condition parameters into the initial Gaussian dot-dyeing model to generate a timing signal, thereby obtaining a corresponding predicted timing signal, wherein the predicted timing signal includes a predicted signal amplitude consistent with the number of reference sampling points; A loss value is calculated based on the predicted signal amplitude and the sampled signal amplitude, and the initial parameters are adjusted based on the loss value until an iteration termination condition is reached, thereby obtaining the trained initial Gaussian dot-dyeing model.

5. A timing signal generating device, characterized in that: include: Data input module: used to obtain time series condition parameters and the number of time series sampling points, and input the time series condition parameters into a pre-trained Gaussian dot-dyeing model, wherein the Gaussian dot-dyeing model includes multiple two-dimensional Gaussian elliptical surfaces; Parameter calculation module: used to determine the scaling vector of each of the two-dimensional Gaussian ellipsoids according to the timing condition parameters, a first scaling factor and a second scaling factor, wherein the first scaling factor and the second scaling factor respectively correspond to the coordinate vector of the ellipse center of the two-dimensional Gaussian ellipsoid; obtain a covariance matrix according to the scaling vector and the transposed vector of the scaling vector, and perform projection based on the coordinate vector of the ellipse center to obtain a corresponding projection vector; calculate the Jacobian matrix between the projection vector and the coordinate vector, use the Jacobian matrix as the projection transformation matrix, obtain a scaling transformation matrix according to the projection transformation matrix and the scaling vector, obtain a projection covariance matrix according to the scaling transformation matrix, the covariance matrix and the transposed matrix of the scaling transformation matrix, obtain the target transparency of the two-dimensional Gaussian ellipsoid, and calculate the difference vector between the projection vector and the coordinate vector, and obtain the projection transparency based on the target transparency, the difference vector and the inverse matrix of the projection covariance matrix; A signal generation module is configured to generate a number of timing sampling points that is consistent with the number of the timing sampling points, calculate at least one target Gaussian ellipse corresponding to each of the timing sampling points, sort the target Gaussian ellipse surfaces according to the projection transparency, calculate the signal amplitude corresponding to the timing sampling points according to the sorting result, and obtain the target timing signal based on the signal amplitude.

6. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the timing signal generation method according to any one of claims 1 to 4 when executing the computer program.

7. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the timing signal generating method according to any one of claims 1 to 4 is implemented.

Images