An over-the-horizon chirp spread spectrum communication signal-to-noise ratio estimation method, system and storage medium

Abstract

The application relates to an over-the-horizon chirp spread spectrum communication signal-to-noise ratio estimation method, a system and a storage medium, and relates to the technical field of over-the-horizon communication. The method comprises the following steps: S1, acquiring a received signal in over-the-horizon chirp spread spectrum communication, performing despread and Fourier transform on the received signal to obtain corresponding frequency spectrum data; S2, extracting a useful signal amplitude and a noise signal amplitude from the frequency spectrum data; S3, calculating a log-domain amplitude ratio feature of the received signal by using the useful signal amplitude and the noise signal amplitude; and S4, inputting the log-domain amplitude ratio feature into a pre-trained signal-to-noise ratio estimation model to obtain a signal-to-noise ratio estimation value of the received signal; wherein the pre-trained signal-to-noise ratio estimation model is a piecewise linear regression model trained based on training samples under different signal-to-noise ratio conditions. Compared with the prior art, the application can avoid increasing the calculation overhead of additional discrete Fourier transform, thereby reducing the calculation complexity and accurately calculating the signal-to-noise ratio under different intensity noise environments, and the real-time performance and the estimation accuracy are considered.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of over-the-horizon communication, in particular to an over-the-horizon chirp spread spectrum communication signal-to-noise ratio estimation method, system and storage medium. BACKGROUND

[0002] In over-the-horizon communication, the system needs to accurately estimate the signal-to-noise ratio (SNR) under extremely weak signals and strong background noise to monitor the link quality and optimize the system performance. The commonly used SNR estimation method is the odd-even frequency point method, which uses the characteristic that the theoretical value of the even index frequency point is zero after the periodic repetition sequence is subjected to discrete Fourier transform to estimate the SNR. Specifically, the noise power is estimated by calculating the statistical variance of the even index frequency point, the total signal power (including useful signal and noise) is estimated by calculating the statistical variance of the odd index frequency point, and then the channel SNR is estimated. The above method needs to calculate the mean square energy of all samples, and the computational complexity is high. Since the noise power and the total signal power are calculated independently through odd and even frequency points, the randomness of the noise in the extremely weak signal and strong background noise environment easily causes the noise estimated power to be higher than the total signal estimated power, at this time the signal estimated power is negative, which causes the SNR to be unable to be calculated, therefore the reliability of this kind of method is low. SUMMARY

[0003] Therefore, it is necessary to provide an over-the-horizon chirp spread spectrum communication signal-to-noise ratio estimation method, system and storage medium which can avoid increasing the computational overhead of additional discrete Fourier transform, thereby reducing the computational complexity and accurately calculating the signal-to-noise ratio in different intensity noise environments, and taking into account the real-time performance and estimation accuracy.

[0004] To solve the above technical problems, the technical solutions of the present application are as follows:

[0005] In a first aspect, an over-the-horizon chirp spread spectrum communication signal-to-noise ratio estimation method comprises:

[0006] S1: obtaining a received signal in over-the-horizon chirp spread spectrum communication, performing despreading and Fourier transform on the received signal to obtain corresponding frequency spectrum data;

[0007] S2: extracting a useful signal amplitude and a noise signal amplitude from the frequency spectrum data, wherein the useful signal amplitude is the maximum amplitude of the frequency spectrum data, and the noise signal amplitude is the average amplitude in a predetermined frequency range outside the frequency point corresponding to the useful signal amplitude;

[0008] S3: calculating the log domain amplitude ratio feature of the received signal using the useful signal amplitude and the noise signal amplitude;

[0009] S4: inputting the log domain amplitude feature into a pre-trained signal-to-noise ratio estimation model to obtain a signal-to-noise ratio estimation value of the received signal; wherein the pre-trained signal-to-noise ratio estimation model is a piecewise linear regression model trained based on training samples under different signal-to-noise ratio conditions.

[0010] In a second aspect, a signal-to-noise ratio estimation system for over-the-horizon chirp spread spectrum communication utilizes the above-mentioned method for estimating the signal-to-noise ratio of over-the-horizon chirp spread spectrum communication, and comprises:

[0011] A signal acquisition module is configured to acquire a received signal in over-the-horizon chirp spread spectrum communication, perform de-spreading and Fourier transform on the received signal, and obtain corresponding frequency spectrum data.

[0012] A data extraction module is configured to extract a useful signal amplitude and a noise signal amplitude from the frequency spectrum data, wherein the useful signal amplitude is the maximum amplitude of the frequency spectrum data, and the noise signal amplitude is the average amplitude in a predetermined frequency range outside the frequency point corresponding to the useful signal amplitude.

[0013] A data processing module is configured to calculate a log domain amplitude feature of the received signal by using the useful signal amplitude and the noise signal amplitude.

[0014] A signal-to-noise ratio estimation module is configured to input the log domain amplitude feature into a pre-trained signal-to-noise ratio estimation model to obtain a signal-to-noise ratio estimation value of the received signal; wherein the pre-trained signal-to-noise ratio estimation model is a piecewise linear regression model trained based on training samples under different signal-to-noise ratio conditions.

[0015] In a third aspect, a computer readable storage medium stores at least one instruction, at least one program, a code set or an instruction set, which are loaded and executed by a processor to implement the above-mentioned method for estimating the signal-to-noise ratio of over-the-horizon chirp spread spectrum communication.

[0016] Compared with the prior art, the technical scheme of the present application has the following beneficial effects:

[0017] The present application combines data driving and chirp spread spectrum waveform features, estimates SNR by using frequency spectrum data generated during over-the-horizon communication demodulation, avoids increasing the calculation overhead of additional discrete Fourier transform, does not need mean square operation compared with the odd-even frequency point method, thereby reducing the calculation complexity, and balances the real-time performance and estimation accuracy; the model optimization is realized based on data driving, and the SNR under different intensity noise environments can be accurately calculated, and the estimation accuracy is greatly improved compared with the traditional odd-even frequency point method in a weak signal strong noise environment. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 A flowchart of a method for estimating a signal-to-noise ratio of an over-the-horizon chirp spread spectrum communication in some embodiments of the present application;

[0019] Figure 2 A flowchart of a method for training a piecewise linear regression model in a method for estimating a signal-to-noise ratio of an over-the-horizon chirp spread spectrum communication in some embodiments of the present application;

[0020] Figure 3 A frequency spectrum data diagram of a despread signal after a discrete Fourier transform in an additive white Gaussian noise channel with an SNR of -20 dB in some embodiments of the present application;

[0021] Figure 4 A performance comparison diagram of a method for estimating a signal-to-noise ratio of an over-the-horizon chirp spread spectrum communication and a mainstream energy-based odd-even frequency point method in some embodiments of the present application;

[0022] Figure 5 A flowchart of a system for estimating a signal-to-noise ratio of an over-the-horizon chirp spread spectrum communication in some embodiments of the present application. DETAILED DESCRIPTION

[0023] The terms "first", "second", and the like in the description and in the claims of the present application are used for distinguishing between similar objects talking about the embodiments of the application and do not necessarily have to appear in the description and claims of the present application in this particular order. It is to be understood that the terms so used are interchangeable under appropriate circumstances and embodiments of the application are meant to encompass all combinations given for all of the terms. Additionally, the term "comprising" is used herein to mean that the embodiments of the application include, but are not limited to, the following steps, elements, components, or options as listed in the description and / or claims of the present application, and that steps, elements, components or options can be added and / or deleted according to the application without departing from the spirit and scope of the present application. Further, the term "determining" encompasses a wide variety of actions and, therefore, the term "determining" can include calculating, computing, processing, deriving, investigating, looking up (such as looking up in a table, a database or another data structure), ascertaining and the like. Also, the term "determining" can include receiving (such as receiving information), accessing (such as accessing data in a memory) and the like. Furthermore, the term "determining" can include resolving, selecting, choosing, establishing and the like.

[0024] It is to be understood that the hereinto described components, elements and / or structures can also be provided in other embodiments of the application in an inverted order. In addition, it is to be understood that the terms "connected to", "coupled to" and the like, refer not only to direct connections, but also to indirect connections via another component, element, structure, etc. between the elements that are "connected to", "coupled to" and the like.

[0025] It should be emphasized that the acquisition, transmission, storage, use, processing, etc. of data in the technical solutions of the embodiments of the present application comply with the relevant provisions of national laws and regulations.

[0026] In the embodiments of the present application, some existing industry solutions may be mentioned, such as certain software, components, models, etc. They should be considered as exemplary, and the purpose is only to illustrate the feasibility of the implementation of the technical solutions of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0027] The accompanying drawings are only used for illustrative purposes and cannot be understood as a limitation of the patent;

[0028] In order to better illustrate the embodiments, some components in the drawings may be omitted, enlarged or reduced, and do not represent the actual size of the product;

[0029] For those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0030] The technical solutions of the present application will be further described below in conjunction with the drawings and embodiments.

[0031] Embodiment 1

[0032] The present embodiment provides a method for estimating the signal-to-noise ratio of over-the-horizon chirp spread spectrum communication, referring to Figure 1 , comprising:

[0033] S1: obtaining a received signal in over-the-horizon chirp spread spectrum communication, performing despreading and Fourier transform on the received signal to obtain corresponding frequency spectrum data;

[0034] S2: extracting a useful signal amplitude and a noise signal amplitude from the frequency spectrum data, wherein the useful signal amplitude is the maximum amplitude of the frequency spectrum data, and the noise signal amplitude is the average amplitude in a predetermined frequency range outside the frequency point corresponding to the useful signal amplitude;

[0035] S3: calculating the log domain amplitude ratio feature of the received signal using the useful signal amplitude and the noise signal amplitude;

[0036] S4: inputting the log domain amplitude ratio feature into a pre-trained signal-to-noise ratio estimation model to obtain a signal-to-noise ratio estimation value of the received signal; wherein the pre-trained signal-to-noise ratio estimation model is a piecewise linear regression model trained based on training samples under different signal-to-noise ratio conditions.

[0037] Compared with the prior art, the method combines data driving and chirp spread spectrum waveform characteristics, estimates SNR by using spectrum data generated during over-the-horizon communication demodulation, can avoid increasing the calculation overhead of additional discrete Fourier transform, does not need mean square operation compared with the odd-even frequency point method, thereby reducing the calculation complexity, and balances the real-time performance and the estimation accuracy; the model optimization is realized based on data driving, SNR in different intensity noise environments can be accurately calculated, and the estimation accuracy is greatly improved compared with the traditional odd-even frequency point method in a weak signal strong noise environment.

[0038] In chirp spread spectrum communication, the despread signal presents a single tone characteristic, and the SNR is estimated based on the characteristic, and it is necessary to perform mathematical derivation, and the specific process is as follows:

[0039] Let the difference between the highest frequency and the lowest frequency in chirp linear spread spectrum modulation be bandwidth . Basically, the frequency of the uplink chirp signal increases linearly with time, the starting frequency is , and the ending frequency is , as shown in equation (1); basically, the frequency of the downlink chirp signal decreases linearly with time, the starting frequency is , and the ending frequency is , as shown in equation (2). The transmitted data information is modulated as the offset of the starting frequency of the uplink chirp signal, which is called the uplink chirp signal.

[0040] (1)

[0041] (2)

[0042] The signal bandwidth is divided into orthogonal chips (the spreading factor is ), and the frequency interval is . Each modulation symbol is composed of chips, which can carry bits of information, and corresponds to different frequency offsets. For symbol , the starting frequency offset is , and the starting frequency of the corresponding modulation signal is . The symbol period is , and the chip duration is . The mathematical expression of the uplink chirp signal can be expressed as:

[0043] (3)

[0044] wherein is a step function, .

[0045] In the discrete-time model, let the sampling rate be and satisfy The sampling interval , i.e., one sample per chip, is called single sampling, and the expression of equation (3) at time can be expressed as:

[0046] (4)

[0047] where , .

[0048] The discrete-time expressions of the basic up-chirp signal corresponding to equation (1) and the basic down-chirp signal corresponding to equation (2) are as follows:

[0049] (5)

[0050] (6)

[0051] In the case of an additive white Gaussian noise (AWGN) channel, the received signal , where represents Gaussian white noise. Using a non-correlation demodulation method, the received signal is first multiplied by the basic down-chirp signal to complete the despreading of the received signal, and a single-tone signal is obtained. The expression of the despreading signal is as follows:

[0052] (7)

[0053] Next, the despreading signal is subjected to FFT operation to convert the time domain to the frequency domain; finally, the discrete frequency index that maximizes is found, which is the data corresponding to the signal code, where the frequency point corresponds to the signal frequency point, and the rest are noise frequency points:

[0054] (8)

[0055] This method uses the above characteristics to extract the log amplitude ratio of the despreading spectrum in the over-the-horizon waveform as an input feature, and based on data driving, a higher-precision SNR estimation is completed through a piecewise linear regression algorithm.

[0056] Embodiment 2

[0057] This embodiment further provides an over-the-horizon chirp spread spectrum communication SNR estimation method based on embodiment 1, comprising:

[0058] S1: obtaining a received signal in over-the-horizon chirp spread spectrum communication, performing despreading and Fourier transform on the received signal to obtain corresponding frequency spectrum data;

[0059] S2: extracting a useful signal amplitude and a noise signal amplitude from the frequency spectrum data, wherein the useful signal amplitude is the maximum amplitude of the frequency spectrum data, and the noise signal amplitude is the average amplitude in a predetermined frequency range outside the frequency point corresponding to the useful signal amplitude;

[0060] S3: calculating a log domain amplitude ratio feature of the received signal using the useful signal amplitude and the noise signal amplitude;

[0061] S4: inputting the log domain amplitude ratio feature into a pre-trained signal-to-noise ratio estimation model to obtain a signal-to-noise ratio estimation value of the received signal; wherein the pre-trained signal-to-noise ratio estimation model is a piecewise linear regression model trained based on training samples under different signal-to-noise ratio conditions.

[0062] In some preferred embodiments, in step S4, the piecewise linear regression model includes a plurality of subset linear regression models, each of which corresponds to a preset signal-to-noise ratio interval;

[0063] Training the piecewise linear regression model to be trained using the training samples, referring to Figure 2 , including:

[0064] According to the signal-to-noise ratio interval, the training sample set is divided into a plurality of subsets, and each subset linear regression model is trained using each subset, and the least square algorithm is used to obtain the optimal parameter estimation of each subset linear regression model corresponding to each signal-to-noise ratio interval using the mean square error as the loss function, to obtain the trained piecewise linear regression model;

[0065] Wherein, by obtaining historical received signals under different signal-to-noise ratio conditions, performing despreading and Fourier transform on the historical received signals to obtain corresponding training frequency spectrum data;

[0066] Extracting a training useful signal amplitude and a training noise signal amplitude from each of the training frequency spectrum data, calculating a log domain amplitude ratio feature of each of the historical received signals; taking the signal-to-noise ratio as the true label of the historical received signal, adding noise to the data when constructing the training sample, and recording the signal-to-noise ratio at this time; constructing the training sample using the calculated log domain amplitude ratio feature.

[0067] In some preferred embodiments, the useful signal amplitude and the noise signal amplitude are extracted from the spectrum data, and the training useful signal amplitude and the training noise signal amplitude are extracted from the training spectrum data, according to the following formula:

[0068]

[0069]

[0070]

[0071] wherein, represents the noise signal amplitude or the training noise signal amplitude, represents the useful signal amplitude or the training useful signal amplitude, represents a noise signal sampling window size, represents a discrete Fourier transform of signal despread data, represents a useful signal frequency point, is a frequency interval between a sampling window starting frequency and , used to eliminate the energy spread effect of the useful signal on the noise amplitude estimation.

[0072] The over-the-horizon waveform has an extremely low demodulation threshold, and the obtained based on the despread spectrum of the standard uplink chirp signal in the linear frequency modulation wave can accurately reflect the energy ratio relationship under the condition of strong noise and weak signal, and ensure accurate estimation of SNR under extremely low SNR. Figure 3 The spectrum data after the discrete Fourier transform of the despread signal under the condition of additive white Gaussian noise channel with SNR of-20dB is shown.

[0073] In some preferred embodiments, a logarithmic domain transformation is performed on the training useful signal amplitude and the training noise signal amplitude to obtain a logarithmic domain amplitude ratio feature of each of the historical received signals, and the is input to the segmented linear regression model to be trained, wherein, represents the signal-to-noise ratio, respectively represents the start point and the end point of the interval of the signal-to-noise ratio;

[0074] In this embodiment, the is used to obtain a logarithmic domain amplitude ratio feature of the current signal. Since , is not equal to the power of the real signal, the logarithmic domain amplitude ratio feature is not equal to the SNR, and the SNR needs to be estimated through feature transformation of the machine learning model. Directly inputting With Mapping back to SNR increases the complexity of the model, which is not conducive to data fitting and hardware implementation. By extracting the log domain features through data preprocessing, the model input can be aligned with the label space, which helps to reduce the training difficulty and deployment difficulty of the model, and improve the estimation accuracy.

[0075] Based on a plurality of separate breakpoints According to the interval of the signal-to-noise ratio, the signal-to-noise ratio is divided into a plurality of subsets , wherein , wherein The first subset is represented by the first subset , wherein The first subset is represented by the first subset , wherein There are .

[0076] In this embodiment, the entire training sample SNR change range is from-21dB to +30dB, covering a variety of intensity noise environments, and each SNR interval contains 200 groups of samples to ensure the richness of the training data.

[0077] In the specific implementation process, the number of separate breakpoints The number of segmented linear regression models used in the application is set to 14.

[0078] In some preferred embodiments, according to the plurality of subsets and the plurality of preset signal-to-noise ratio intervals corresponding thereto, the segmented linear regression model to be trained is constructed, which is represented as:

[0079]

[0080] , wherein And Respectively represent the bias parameter and the weight parameter of the first subset linear regression model in the linear regression model , wherein The signal-to-noise ratio fitting target value is

[0081] In some preferred embodiments, according to the The number of And the corresponding real label of the first subset is used to construct the training sample set , wherein The real label corresponding to the first subset The number of The real label corresponding to the first subset is represented by the real label corresponding to the first subset The number of According to the training sample set, each subset linear regression model is independently trained;

[0082] ​The mean square error loss function is represented as:

[0083]

[0084]

[0085]

[0086]

[0087] wherein, is the value of the mean square error loss function; , and is the sample matrix corresponding to the kth subset, is the parameter vector of the subset linear regression model corresponding to the kth subset, and is the label vector corresponding to the kth subset;

[0088] The optimal parameter estimation of the subset linear regression model corresponding to each signal-to-noise ratio interval is obtained by using a least square algorithm , and is represented as:

[0089]

[0090] wherein, represents a conjugate transpose.

[0091] In some preferred embodiments, the pre-trained signal-to-noise ratio estimation model is represented as:

[0092]

[0093] wherein, is the signal-to-noise ratio estimation value of the received signal; according to a plurality of the separation breakpoints , the boundary of the log domain amplitude ratio of the kth subset linear model in the pre-trained signal-to-noise ratio estimation model is determined ; ;

[0094] According to the size relationship between the log domain amplitude ratio of the received signal and the boundary, the corresponding subset linear regression model is selected, the linear transformation of the log feature to the signal-to-noise ratio is completed by the subset linear regression model, and the signal-to-noise ratio estimation value of the received signal is obtained.

[0095] In some preferred embodiments, the received signal and the historical received signal are both signals with a preamble data.

[0096] The performance of the algorithm of the present application is compared with that of the mainstream energy-based odd-even frequency point method, and reference is made to Figure 4, the vertical axis represents the absolute value of the SNR estimation error, and the horizontal axis represents the true SNR. In the experiment, the average value of 100 random tests at each SNR was used as the estimation value for both algorithms, and any complex estimation was considered as an estimation failure. Figure 4 It can be seen that the algorithm of the present application has obvious advantages compared with the odd-even frequency point method at low SNR. The odd-even frequency point method is prone to the case that the noise energy estimation value is greater than the total energy estimation value at low SNR, resulting in that the SNR is calculated as a complex number under this condition, and the value is invalid. The algorithm of the present application is based on data driving, and the corresponding regression model is optimized for different SNRs, so it shows extremely high estimation accuracy at low SNR and high SNR. This experiment fully proves that the algorithm of the present application can accurately estimate SNR in an extremely weak signal and strong background noise environment, which has important significance for channel monitoring under complex conditions, and is especially suitable for SNR estimation in over-the-horizon communication systems.

[0097] The embodiment also provides an over-the-horizon chirp spread spectrum communication SNR estimation system, which utilizes the over-the-horizon chirp spread spectrum communication SNR estimation method described above, and refers to Figure 5 , and comprises:

[0098] A signal acquisition module is configured to acquire a received signal in over-the-horizon chirp spread spectrum communication, perform de-spreading and Fourier transform on the received signal, and obtain corresponding frequency spectrum data.

[0099] A data extraction module is configured to extract a useful signal amplitude and a noise signal amplitude from the frequency spectrum data, wherein the useful signal amplitude is the maximum amplitude of the frequency spectrum data, and the noise signal amplitude is the average amplitude in a predetermined frequency range outside the frequency point corresponding to the useful signal amplitude.

[0100] A data processing module is configured to calculate a log domain amplitude feature of the received signal by using the useful signal amplitude and the noise signal amplitude.

[0101] An SNR estimation module is configured to input the log domain amplitude feature into a pre-trained SNR estimation model to obtain an SNR estimation value of the received signal, wherein the pre-trained SNR estimation model is a piecewise linear regression model trained based on training samples under different SNR conditions.

[0102] It can be understood that the system of the embodiment corresponds to the method described above, and the optional items in the above method are also applicable to the system, so they will not be described again here.

[0103] Embodiment 3

[0104] The embodiment provides a computer readable storage medium, and at least one instruction, at least one program, a code set or an instruction set are stored on the storage medium. The at least one instruction, the at least one program, the code set or the instruction set are loaded and executed by a processor, so that the processor executes part or all steps of the method provided in the embodiment 2 of the application.

[0105] It can be understood that the storage medium can be transitory or non-transitory. Exemplarily, the storage medium includes but is not limited to a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0106] Exemplarily, the processor can be a central processing unit (CPU), a microprocessor unit (MPU), a digital signal processor (DSP) or a field programmable gate array (FPGA).

[0107] Exemplarily, the read-only memory includes but is not limited to a mask ROM, a PROM, an EPROM, an EEPROM, a flash and the like.

[0108] Exemplarily, the random access memory includes but is not limited to a DRAM, an SRAM, an SDRAM, a DDR SDRAM and the like.

[0109] In some examples, a computer program product is provided, which can be implemented in a hardware, software or combination thereof. As a non-limiting example, the computer program product can be embodied in the storage medium, and can also be embodied in a software product, such as an SDK (Software Development Kit) and the like.

[0110] As a non-limiting example, a computer program product is provided, which includes a computer program or computer executable instructions stored in a computer readable storage medium. A processor of an electronic device reads the computer program or computer executable instructions from the computer readable storage medium, and the processor executes the computer executable instructions, so that the electronic device executes part or all steps of the method provided in the embodiment of the application.

[0111] In some examples, a computer program is provided, including computer readable code, which, when run in a computer device, causes a processor in the computer device to perform some or all of the steps in the method.

[0112] The embodiment also provides an electronic device, including a memory and a processor, the memory stores at least one instruction, at least one program, a code set or an instruction set, and the processor implements some or all of the steps of the method as described in embodiment 1 when executing the at least one instruction, at least one program, code set or instruction set.

[0113] In some examples, a hardware entity of the electronic device is provided, including a processor, a memory and a communication interface; wherein the processor generally controls the overall operation of the electronic device; the communication interface is used for the electronic device to communicate with other terminals or servers through a network; the memory is configured to store instructions and applications executable by the processor, and can also cache data to be processed by the processor and data to be processed or having been processed by each module in the electronic device (including but not limited to image data, audio data, voice communication data and video communication data), which can be realized by FLASH, EPROM (Erasable Programmable Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory) or RAM (Random Access Memory).

[0114] The processor can include one or more processing elements. Thus, the processor can include one or more integrated circuits (ICs) that are configured to perform the functions of the processor. In addition, each integrated circuit can include circuitry (e.g., first circuitry, second circuitry, and other circuitry) that is configured to perform the functions of the processor.

[0115] Further, the processor, the communication interface and the memory can transmit data through a bus, which can include any number of interconnected buses and bridges, connecting the various circuitries of the one or more processors and the memory together.

[0116] It can be understood that the optional items in the above embodiment 2 are also applicable to the present embodiment, and thus will not be repeatedly described here.

[0117] The same or similar reference numerals correspond to the same or similar components;

[0118] The terms describing the positional relationship in the drawings are only used for exemplary illustration, and cannot be understood as a limitation on the present application;

[0119] It should be noted that the embodiments and features of the embodiments in the present application can be combined with each other without conflict.

[0120] In different specific implementations, the method or system described in the present application can be implemented in software, hardware or a combination thereof. In addition, the order of the steps of the method can be changed, and various elements can be added, reordered, combined, omitted, modified, etc.

[0121] Obviously, the above embodiments of the present application are only examples for clearly illustrating the present application, and are not intended to limit the implementation of the present application, and are not used to limit the present application. Based on the above description, other different forms of changes or variations can be made by those of ordinary skill in the art, and each separate structure / function module or unit can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part, and the structure and function of the separate components can be implemented as a combined structure or component. Here, it is not necessary and impossible to enumerate all the implementations. Any modification, equivalent replacement and improvement made within the spirit and principle of the present application shall be included in the protection scope of the claims of the present application.

Claims

1. A method for estimating the signal-to-noise ratio (SNR) of beyond-line-of-sight chirp spread spectrum communication, characterized in that, include: S1: Obtain the received signal in the beyond-line-of-sight chirp spread spectrum communication, and perform despreading and Fourier transform on the received signal to obtain the corresponding spectrum data; S2: Extract the useful signal amplitude and noise signal amplitude from the spectrum data, wherein the useful signal amplitude is the maximum amplitude of the spectrum data, and the noise signal amplitude is the average amplitude within a predetermined frequency range outside the frequency point corresponding to the useful signal amplitude; S3: Calculate the logarithmic domain amplitude ratio characteristic of the received signal using the amplitude of the useful signal and the amplitude of the noise signal; S4: Input the logarithmic domain amplitude ratio feature into the pre-trained signal-to-noise ratio estimation model to obtain the signal-to-noise ratio estimate of the received signal; wherein, the pre-trained signal-to-noise ratio estimation model is a piecewise linear regression model trained based on training samples under different signal-to-noise ratio conditions; The piecewise linear regression model includes several subset linear regression models, each of which corresponds to a preset signal-to-noise ratio interval; Training the piecewise linear regression model to be trained using the training samples includes: The training sample set is divided into multiple subsets according to the signal-to-noise ratio interval. The corresponding subset linear regression model is trained using each subset. The mean squared error is used as the loss function. The least squares algorithm is used to obtain the optimal parameter estimate of the subset linear regression model corresponding to each signal-to-noise ratio interval, and the piecewise linear regression model is obtained after training. Specifically, by acquiring historical received signals under different signal-to-noise ratio conditions, the historical received signals are despread and Fourier transformed to obtain the corresponding training spectrum data; The amplitude of the useful training signal and the amplitude of the noise training signal are extracted from each of the training spectrum data, and the logarithmic domain amplitude ratio feature of each of the historical received signals is calculated respectively; the signal-to-noise ratio is used as the true label of the historical received signal; and the training sample is constructed using the calculated logarithmic domain amplitude ratio feature.

2. The signal-to-noise ratio estimation method for beyond-line-of-sight chirp spread spectrum communication according to claim 1, characterized in that, The useful signal amplitude and the noise signal amplitude are extracted from the spectrum data according to the following formulas, and the useful signal amplitude for training and the noise signal amplitude for training are extracted from the training spectrum data: in, This refers to the amplitude of the noise signal or the amplitude of the training noise signal. This refers to the amplitude of the useful signal or the amplitude of the useful signal used for training. Indicates the size of the noise signal sampling window. The discrete Fourier transform of the despread signal data is represented. Indicates the frequency of the useful signal. The sampling window start frequency and The frequency interval between them is used to eliminate the influence of the useful signal on the energy diffusion effect when estimating the noise amplitude.

3. The signal-to-noise ratio estimation method for beyond-line-of-sight chirp spread spectrum communication according to claim 2, characterized in that, Logarithmic domain transformation is performed on the amplitudes of the useful training signal and the noise training signal to obtain the logarithmic domain amplitude ratio characteristics of each of the historical received signals. and the The input is fed into the piecewise linear regression model to be trained, wherein, This indicates the signal-to-noise ratio. These represent the start and end points of the signal-to-noise ratio interval, respectively. Based on multiple dividing points Based on the range of the signal-to-noise ratio Divided into multiple subsets ,in, In Indicates the first A subset Indicates the first Individuals are concentrated indivual .

4. The signal-to-noise ratio estimation method for beyond-line-of-sight chirp spread spectrum communication according to claim 3, characterized in that, Based on the multiple subsets and their corresponding multiple preset signal-to-noise ratio intervals, the piecewise linear regression model to be trained is constructed, which is expressed as: in, and They represent the first and second lines in the linear regression model, respectively. The bias parameters and weight parameters of the linear regression model for the subset mentioned. The target value is the signal-to-noise ratio.

5. The signal-to-noise ratio estimation method for beyond-line-of-sight chirp spread spectrum communication according to claim 4, characterized in that, According to the Sub-concentration indivual The training sample set is constructed using the corresponding real labels. ,in, Indicates the corresponding number Sub-concentration indivual The true labels; the linear regression models of each subset are trained independently based on the training sample set; The mean squared error loss function is expressed as: in, This represents the value of the mean squared error loss function; , and These are, respectively, the sample matrix corresponding to the k-th subset, the parameter vector of the linear regression model for the k-th subset, and the label vector corresponding to the k-th subset; The optimal parameter estimates of the subset linear regression model for each signal-to-noise ratio interval are obtained using the least squares algorithm. , is represented as: in, This indicates the conjugate transpose.

6. The signal-to-noise ratio estimation method for beyond-line-of-sight chirp spread spectrum communication according to claim 5, characterized in that, The pre-trained signal-to-noise ratio estimation model is expressed as follows: in, The signal-to-noise ratio estimate of the received signal; based on the multiple said separation points Determine the pre-trained signal-to-noise ratio estimation model. Log-domain magnitude ratio of subset linear models boundary ; Based on the relationship between the logarithmic domain amplitude ratio and the boundary of the received signal, the corresponding subset linear regression model is selected. The subset linear regression model performs a linear transformation from logarithmic features to signal-to-noise ratio, thereby obtaining the signal-to-noise ratio estimate of the received signal.

7. A signal-to-noise ratio estimation method for beyond-line-of-sight chirp spread spectrum communication according to any one of claims 1-6, characterized in that, Both the received signal and the historical received signal are signals with preamble data.

8. A signal-to-noise ratio (SNR) estimation system for beyond-line-of-sight chirp spread spectrum communication, utilizing the SNR estimation method for beyond-line-of-sight chirp spread spectrum communication as described in any one of claims 1 to 7, characterized in that, include: The signal acquisition module is used to acquire the received signal in beyond-line-of-sight chirp spread spectrum communication, and to perform despreading and Fourier transform on the received signal to obtain the corresponding spectrum data. The data extraction module is used to extract the amplitude of a useful signal and the amplitude of a noise signal from the spectrum data, wherein the amplitude of the useful signal is the maximum amplitude of the spectrum data, and the amplitude of the noise signal is the average amplitude within a predetermined frequency range outside the frequency point corresponding to the amplitude of the useful signal; The data processing module is used to calculate the logarithmic domain amplitude ratio characteristic of the received signal using the amplitude of the useful signal and the amplitude of the noise signal; The signal-to-noise ratio (SNR) estimation module is used to input the logarithmic domain amplitude ratio feature into a pre-trained SNR estimation model to obtain the SNR estimate of the received signal; wherein, the pre-trained SNR estimation model is a piecewise linear regression model trained based on training samples under different SNR conditions.

9. A computer-readable storage medium, characterized in that, The storage medium stores at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, at least one program, code set, or instruction set is loaded and executed by a processor to implement the method as described in any one of claims 1-7.

Images