Method for interference quantification analysis of communication systems for cognitive decision

Abstract

The present application relates to a kind of communication system interference quantitative analysis methods for cognitive decision, belong to wireless communication technical field.The method in communication confrontation, the frequency spectrum of interference signal is as input, based on the communication signal of being interfered, using the disturbed analysis model corresponding to the type of interference signal identified, the communication signal of being interfered is quantitatively analyzed.The method can quickly and accurately identify single tone interference, multi-tone interference, narrowband interference, wideband interference, pulse interference or aiming interference etc.The type of interference signal reduces the time of determining disturbed analysis model as a whole, to meet the timeliness requirement in wartime.Through the quantitative analysis of the communication signal of being interfered, for the selection of communication mode, the setting of communication parameter etc.in complex environment in wartime, provide theoretical basis, for the corresponding countermeasures when being interfered, realize the effective anti-interference of communication system and provide decision support.

Description

Technical Field

[0001] This disclosure relates to the field of wireless communication technology, and more particularly to a quantitative analysis method for interference in communication systems used in cognitive decision-making within the field of communication countermeasures technology. Background Technology

[0002] In today's era of rapid information technology development, information dominance plays a decisive role in modern warfare. Therefore, communication jamming techniques aimed at attacking and disrupting information transmission are unavoidable in modern warfare. Future communication systems are likely to face deliberate combinations of jamming from multiple sources and in various patterns. Traditional anti-jamming methods are no longer sufficient to meet the demands of modern communication technology. Therefore, how to implement effective anti-jamming measures based on existing communication anti-jamming theories is a key focus of communication countermeasures research.

[0003] Currently, most communication systems employ digital modulation methods (commonly M-PSK and M-QAM modulation signals). These digital communications possess strong anti-interference and anti-multipath fading capabilities, thus finding widespread application in tactical communications. Studying the performance of these communication signals under targeted interference from natural environments or non-cooperative parties (such as single-tone interference, multi-tone interference, narrowband interference, and broadband interference) is crucial for research on communication anti-interference technology. The paper "Analysis of the Impact of Single-Tone Interference on 16QAM Demodulation Performance" establishes a single-tone interference model for a 16QAM demodulator, analyzes the impact of single-tone interference signal parameters on demodulator performance, draws conclusions on the influence of interference ratio and single-tone signal frequency offset on system interference effectiveness, and provides the system's bit error rate curve and constellation diagrams for different frequency offsets. The paper "Performance Analysis of MQAM in Single-Tone Interference Environment" presents the demodulation performance of MQAM signals in additive white Gaussian noise channels and single-tone interference environments, and provides a general expression for the MQAM symbol error rate under single-tone interference. It was found that when the frequency of the single-tone interference signal is the same as the carrier signal frequency, the symbol error rate increases with the increase of the power of the single-tone interference signal. Existing interference effect analysis methods either ignore the influence of channel noise or use vector diagram methods for performance analysis. While these methods are simple in concept, they cannot provide accurate performance analysis of the anti-interference performance of communication systems. Summary of the Invention

[0004] To address the aforementioned technical problem of the inability to accurately analyze the anti-interference performance of communication systems, the present invention aims to provide a quantitative analysis method for communication system interference for cognitive decision-making. This method applies a deep convolutional neural network model to communication countermeasures to identify the types of communication interference signals. Then, it combines the interference signals to perform interference performance analysis, providing an effective theoretical basis for anti-interference when communication systems are disrupted during wartime.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows.

[0006] On the one hand, this invention proposes a method for quantitative analysis of interference in communication systems for cognitive decision-making, the method comprising the following steps:

[0007] The spectrum of the interference signal is used as input, and the type of interference signal is identified using a pre-defined deep convolutional neural network model.

[0008] Based on the interfered communication signal, a disturbance analysis model corresponding to the identified interference signal type is used to perform quantitative analysis of the interfered communication signal.

[0009] In the above technical solution, deep learning is used to identify the types of interference signals. This enables real-time and accurate identification of various interference signal types, including single-tone interference, multi-tone interference, narrowband interference, broadband interference, pulse interference, and targeted interference. This allows for rapid determination of the interference analysis model to meet wartime timeliness requirements. Quantitative analysis of the interfered communication signals primarily involves the quantitative calculation of performance indicators. This provides a theoretical basis for selecting communication methods and setting communication parameters in complex wartime environments, and offers decision support for taking appropriate countermeasures when interfered with, thus achieving effective anti-interference of the communication system. Performance indicators include bit error rate, latency, latency-bandwidth product, bit error rate, data rate, and noise.

[0010] As a further improvement to the above technical solution, the interference analysis model uses a bandpass filter to filter out noise outside the signal frequency band of the interfered communication signal, modulates it through a modulator, and then filters out high-frequency components of the modulated signal through an ideal low-pass filter. The signal after filtering out high-frequency components is then sampled and decided to obtain a sampling parameter. The error symbol probability is calculated based on the sign of the sampling parameter to obtain the bit error rate of different signals under different interference signals.

[0011] As an improvement to the above technical solution, after obtaining the theoretical bit error rate of the communication signal after interference, the bit error rate can be reduced by adopting one or more of the following methods in combination: changing the modulation method of the communication signal, the signal power, the signal bandwidth, and the data transmission rate.

[0012] In the above technical solution, the deep convolutional neural network model is composed of two convolutional layers, two pooling layers, and three fully connected layers.

[0013] In the above technical solution, the communication signal is a BPSK signal, a QPSK signal, or a 16QAM signal. The interference analysis model quantitatively analyzes the interference performance of BPSK, QPSK, or 16QAM signals under multiple interference signal types, particularly accurately calculating the theoretical bit error rate. This not only provides an effective anti-interference theoretical basis for wartime communication systems when interfered with, but also helps to deeply understand the interference performance of different interference signal types on different modulation methods, and provides assistance for studying the interference performance of other commonly used modulation methods.

[0014] In one implementation, the interfered communication signal is a BPSK signal. When the interference signal type is single-tone interference, the theoretical bit error rate of the BPSK signal is calculated using the following formula:

[0015]

[0016] When the interference signal type is multi-tone interference, the theoretical bit error rate of the BPSK signal is calculated using the following formula:

[0017]

[0018] When the interference signal type is narrowband interference or wideband interference, the theoretical bit error rate of the BPSK signal is calculated using the following formula:

[0019]

[0020] In the formula: P e T represents the bit error rate, A is the amplitude of the BPSK signal, and T is the bit error rate. s Where J is the symbol time width, and J0 is the power of the single-tone interference signal. Let Q be the phase of the single-tone interference signal, and let Q be the right-tail function of the standard normal distribution. Here, L represents the noise variance; L indicates that L interference spectral lines were applied; and J1 is the power of the multitone interference signal. The phase of the multi-tone interference signal, Δf i The frequency interval between the interference signal and the carrier signal is represented by N0; N0 is the noise power spectral density. J3 represents the broadband / narrowband interference power spectral density, P J For broadband / narrowband interference signal power, B J For broadband / narrowband interference signal bandwidth, E b Let sin c(x) be the signal energy averaged over each bit, and let x be the singer function.

[0021] In one implementation, the interfered communication signal is a QPSK signal. When the interference signal type is single-tone interference, the theoretical bit error rate of the QPSK signal is calculated using the following formula:

[0022]

[0023] When the interference signal type is multi-tone interference, the theoretical bit error rate of the QPSK signal is calculated using the following formula:

[0024]

[0025] When the interference signal type is narrowband interference or wideband interference, the theoretical bit error rate of the QPSK signal is calculated using the following formula:

[0026]

[0027] In the formula: P e T represents the bit error rate. s Where J is the symbol time width, and J0 is the power of the single-tone interference signal. Let Q be the phase of the single-tone interference signal, and let Q be the right-tail function of the standard normal distribution. Here, L represents the noise variance; L indicates that L interference spectral lines were applied; and J1 is the power of the multitone interference signal. The phase of the multi-tone interference signal, Δf i S represents the frequency interval between the interference signal and the carrier signal. QPSK J3 = P, where N is the power of the QPSK signal and N0 is the noise power spectral density. J / B J J3 represents the broadband / narrowband interference power spectral density, P J For broadband / narrowband interference signal power, B J For broadband / narrowband interference signal bandwidth, E b Let sin c(x) be the signal energy averaged over each bit, and let x be the singer function.

[0028] In one implementation, the interfered communication signal is a 16QAM signal. When the interference signal type is single-tone interference, the theoretical bit error rate of the 16QAM signal is calculated using the following formula:

[0029]

[0030] in:

[0031]

[0032]

[0033]

[0034]

[0035] When the interference signal type is multi-tone interference, the theoretical bit error rate of the 16QAM signal is calculated using the following formula:

[0036]

[0037] in:

[0038]

[0039]

[0040]

[0041]

[0042] When the interference signal type is narrowband interference or wideband interference, the theoretical bit error rate of the 16QAM signal is calculated using the following formula:

[0043]

[0044] In the formula: P e T represents the bit error rate. s Where J is the symbol time width, and J0 is the power of the single-tone interference signal. Let Q be the phase of the single-tone interference signal, and let Q be the right-tail function of the standard normal distribution. Here, L represents the noise variance; L indicates that L interference spectral lines were applied; and J1 is the power of the multitone interference signal. The phase of the multi-tone interference signal, Δf i The frequency interval between the interference signal and the carrier signal is represented by N0; the noise power spectral density is S. 16QAM J3 = P, representing the power of a 16QAM signal. J / B J J3 represents the broadband / narrowband interference power spectral density, P J For broadband / narrowband interference signal power, B J For broadband / narrowband interference signal bandwidth, E b Let sin c(x) be the signal energy averaged over each bit, and let x be the singer function.

[0045] On the other hand, the present invention proposes a computer-readable storage medium storing a computer program that can be loaded by a processor and executed by any of the methods described above. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 , one A schematic diagram of the method flow of the present invention in one embodiment;

[0048] Figure 2(a) , one Performance analysis diagram of BPSK signal under single-tone interference in one implementation method;

[0049] Figure 2(b) , one Performance analysis diagram of BPSK signal under multi-tone interference in one implementation method;

[0050] Figure 2(c) , one Performance analysis of the BPSK signal under narrowband interference in one implementation method;

[0051] Figure 2(d) , one Performance analysis diagram of BPSK signal under broadband interference in one implementation method;

[0052] Figure 3(a) , one Performance analysis diagram of QPSK signal under single-tone interference in one implementation method;

[0053] Figure 3(b) , one Performance analysis of QPSK signal under multi-tone interference in one implementation method;

[0054] Figure 3(c) , one Performance analysis of QPSK signal under narrowband interference in one implementation method;

[0055] Figure 3(d) , one Performance analysis diagram of QPSK signal under broadband interference in one implementation method;

[0056] Figure 4(a) , one Performance analysis diagram of 16QAM signal under single-tone interference in one implementation method;

[0057] Figure 4(b) , one Performance analysis diagram of 16QAM signal under multi-tone interference in one implementation method;

[0058] Figure 4(c) , one Performance analysis of the 16QAM signal under narrowband interference in one implementation method;

[0059] Figure 4(d) , one Performance analysis diagram of 16QAM signal under broadband interference in one implementation method;

[0060] Figure 5 , one The accuracy of identifying different types of interference signals in this implementation method. Detailed Implementation

[0061] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0062] In one implementation, after conducting communication reconnaissance against the enemy and acquiring the technical parameters of the jamming signal, a deep reinforcement learning method is used to identify the type of jamming signal. Based on known signal parameters such as the modulation scheme, symbol rate, and carrier frequency of the friendly communication signal, as well as information such as the type and frequency of the jamming signal applied by the enemy, a formula is used to calculate the theoretical bit error rate (BER). Then, the BER variation of the communication signal under different signal-to-interference ratios is analyzed, providing a basis for effective anti-jamming of wartime communication systems. Based on the analysis results, corresponding anti-jamming measures are taken. The specific implementation includes three parts: jamming signal type identification, jamming signal interference performance analysis, and taking effective anti-jamming measures. (See the schematic diagram.) Figure 1 The flowchart is illustrative and may contain operations that are not performed in order, may be performed in reverse order or simultaneously, and may also include one or more additional operations or remove one or more operations from the flowchart.

[0063] 1. Identification of interference signal types

[0064] (1.1) Obtaining Interference Signal Parameters

[0065] For the acquired complex interference signals, parameters are extracted, including modulation scheme, symbol rate, and carrier frequency.

[0066] (1.2) Interference signal preprocessing

[0067] Data preprocessing is used to improve the learning speed of the interference signal identification network without affecting the interference signal classification characteristics. The baseband signal x(n) is processed to perform power normalization, specifically as follows:

[0068]

[0069] Where N is the number of signal sampling points.

[0070] The preprocessed data is used to obtain its spectrum, which is then used as input data to the designed interference signal type identification network model.

[0071] (1.3) Design of an interference signal type identification network model

[0072] An interference signal type recognition network was built based on a CNN network. CNNs are among the most widely used networks in image recognition and classification. They are feedforward networks that use convolutional neural networks to extract local feature information from input data. The neurons in different layers of a CNN are not fully connected, so scaling or translation of the data does not change its essential characteristics.

[0073] The interference signal type recognition network consists of two convolutional layers, two pooling layers, and three fully connected layers. The convolutional and pooling layers work together to form multiple convolutional groups, extracting data features layer by layer. Functionally, the network can be divided into two parts: a feature extraction module and a result processing module. The feature extraction module has four layers: 576 nodes in the input layer and 24,576 nodes in the output layer. The result processing module contains three fully connected layers with 1024, 512, and 6 neurons respectively.

[0074] (1.4) Feed the data into the network for training.

[0075] When training a network for identifying interference signals, gradient descent is used for parameter optimization. Typically, training a convolutional neural network (CNN) requires focusing on three key aspects: weight initialization, forward propagation, and backpropagation. Inappropriate initial weights can lead to slow or even non-convergent convergence during training. Therefore, pre-training can be used to initialize weight parameters, accelerating model convergence and improving generalization ability. The input data, after passing through each layer's corresponding algorithm, is passed to the next layer; this process is called forward propagation. Different layers in the network use different forward propagation algorithms. If the current layer is a convolutional layer, the output after forward propagation is:

[0076] y l =f(y l-1 ×W l +b l )

[0077] Where: y l-1 This is the output of layer (l-1), W l For the convolution kernel, b l Here, f(x) is the bias, and f(x) is the activation function.

[0078] If the current layer is a pooling layer, the output after the forward propagation algorithm is:

[0079] y l =pool(y l-1 )

[0080] Where pool(x) is the pooling operation.

[0081] In addition, the training process includes two operations: error calculation and parameter update. The output result is compared with the expected result using the input data transmitted in the forward pass, and the error E is calculated using the following cross-entropy loss function:

[0082]

[0083] Where: k is the dimension of the data, t k These are the expected result labels, represented using one-hot encoding. The correct solution is labeled with a value of 1, while all other solutions are labeled with a value of 0. k This indicates the output of the network.

[0084] The network parameters are then adjusted using an optimization algorithm; this process is called backpropagation. When the network training error shows a decreasing trend, initially decreasing rapidly but gradually slowing down with each training iteration until it approaches zero, it indicates the end of network training. In practice, training can be stopped by controlling the number of training iterations or by ensuring the error meets a set error threshold.

[0085] The six types of interference signals are clearly distinguishable in the spectrum diagram. Through training a convolutional neural network, the types of interference signals can be effectively identified.

[0086] (1.5) Interference signal type identification

[0087] The trained interference signal recognition network can be used as a preset network model. After inputting the spectrum of the interference signal, the type of interference signal can be directly obtained. The types of interference signals that can be identified include single-tone interference, multi-tone interference, narrowband interference, broadband interference, pulse interference, and targeted interference.

[0088] 2. Establish a disturbance analysis model

[0089] The performance metrics in this invention include bit error rate, latency, latency-bandwidth product, bit error rate, data rate, and noise. In this embodiment, bit error rate is used as an example of a performance metric; the analysis of other performance metrics is similar.

[0090] First, based on different types of interference signals, a corresponding interference analysis model is established, which is based on the modulation method, symbol rate, carrier frequency and other signal parameters of the interfered communication signal and the identified type of interference signal to obtain the theoretical bit error rate of the communication signal after interference.

[0091] Specifically, the disturbance analysis model uses a bandpass filter to remove noise outside the signal's frequency band from the interfered communication signal, modulates the signal using a modulator, and then passes the modulated signal through an ideal low-pass filter to remove high-frequency components. The signal after high-frequency component removal is then sampled and decided. Based on the symbols obtained from the sampling decisions, the probability of erroneous symbols is calculated to obtain the bit error rate (BER) of different signals under different interference conditions. This disturbance analysis model can also be further improved to a model capable of simultaneously performing quantitative analysis of multiple performance indicators.

[0092] In this embodiment, the types of interference signals identified are single-tone interference, multi-tone interference, narrowband interference, and wideband interference. The changes in bit error rate after BPSK, QPSK, and 16QAM signals are obtained after interference are obtained.

[0093] Typically, single-tone interference emits a sine wave at only one frequency. The interference frequency is usually the same as the carrier frequency of the interfered signal, and the power spectrum of the single-tone signal is a single spectral line at the interference frequency. (Single-tone interference J) single The expression for (t) is:

[0094] J single (t)=J0cos(w c t+θ J0 )

[0095] in, It is the phase shift between the interference signal and the carrier signal, J0 is the power of the single-tone interference signal, the frequency of the interference signal is aligned with the frequency of the carrier signal, and w c Let ω be the angular frequency of the signal, and t be time.

[0096] Multi-tone interference signals are generated by the superposition of L independent sinusoidal signal waveforms, and the multi-tone interference J multi The expression for (t) is:

[0097]

[0098] Where: L represents the application of L interfering spectral lines, w i Let w be the angular frequency of the interference signal for the i-th interference spectral line, i = 1, 2, ..., L, and w i Satisfy w i =2πΔf+w c Δf is the frequency interval of the interference signal. J1 is the total power of the multi-tone interference signal.

[0099] Broadband interference is a type of interference signal with a wide bandwidth in the frequency domain. Compared to narrowband interference, it overlaps with the useful signal in the frequency domain by a greater extent, and therefore has a greater impact on the useful signal.

[0100] Narrowband interference, another common type of suppression interference, has a distinct characteristic: its power, relative to the signal, is concentrated within a narrow bandwidth, typically occupying no more than 10% of the bandwidth. In the subsequent simulation, the specific steps for generating narrowband interference are: first, Gaussian white noise is generated; then, a specific narrowband filter is used to filter the Gaussian white noise, and the resulting signal is the desired interference.

[0101] Given the interference method and known signal parameters such as modulation scheme, symbol rate, and carrier frequency, the theoretical bit error rate (BER) formula for the signal after interference can be derived by analyzing the signal modulation and demodulation process. The MATLAB simulation of the signal interference process yields a simulated BER curve, which is then compared with the theoretical formula. By analyzing the potential changes in the communication signal before and after different types of interference, targeted anti-interference measures can be implemented, such as stopping or switching to another channel to achieve anti-interference effects.

[0102] The following section describes the performance changes of BPSK, QPSK, and 16QAM signals under single-tone, multi-tone, narrowband, and wideband interference, and derives their theoretical bit error rate formulas under these interference conditions. Based on the interference analysis models for BPSK, QPSK, and 16QAM, the transmitted signal is always subject to noise interference during channel transmission. By setting a bandpass filter at the front end of the demodulator to filter out noise outside the signal's frequency band, the impact of noise can be reduced.

[0103] Assuming our communication signal is e(t), and the noise in the channel is additive white Gaussian noise, denoted by n(t), after adding external interference J(t), the signal y(t) at the modulator input can be expressed as:

[0104] y(t) = e(t) + n(t) + J(t)

[0105] The modulated signal is then passed through an ideal low-pass filter to remove high-frequency components, and the signal after high-frequency component removal is sampled and decided. Based on the symbols after sampling and decision, the probability of erroneous symbols is calculated, and the bit error rate formula for different signals under different interferences can be obtained.

[0106]

BPSK Disturbance Analysis Model

[0107] (1.1) Theoretical bit error rate of BPSK under single-tone interference

[0108] The noise in the channel is additive white Gaussian noise. After passing through a bandpass filter, the output signal of the filter is narrowband white Gaussian noise n. z (t), at this time:

[0109]

[0110] Where: n c (t) represents the in-phase component of the noise, w c Let ω be the angular frequency of the signal, t be time, and n be the angular frequency. s (t) represents the orthogonal components of the noise. The input signal y(t) at the receiver can be expressed as:

[0111] y(t)=e BPSK (t)+n(t)+J single (t)

[0112] In the formula: e BPSK (t) is the BPSK signal, J single (t) represents single-tone interference.

[0113] The input signal y(t) multiplied by the carrier wave is expressed as:

[0114]

[0115] The output signal x(t) after the signal is filtered out of high-frequency components by an ideal low-pass filter is expressed as:

[0116]

[0117] In the formula: A is the amplitude of the BPSK signal, and J0 is the power of the single-tone interference signal. For the phase of the single-tone interference signal, To represent the phase of the BPSK signal, Or π is the instantaneous phase of the signal, n c (t) represents the in-phase component of the noise.

[0118] The signal after filtering out high-frequency components is sampled and decided, and the sampling parameter x is in the i-th symbol time interval:

[0119]

[0120] Where: a i equal when When π or a i =1 or -1. T represents the integral of the orthogonal components of the noise over the i-th symbol time interval. s This represents the code element time width.

[0121] The error rate P of the BPSK signal under single-tone interference is obtained by calculating the error probability of x based on the sign of x. e The calculation formula is as follows:

[0122] P e =P(x<0|a i=1)+P(x>0|a i =-1)

[0123] in:

[0124]

[0125]

[0126] (1.2) Theoretical bit error rate of BPSK under multi-tone interference

[0127] Input signal: y(t) = e BPSK (t)+n(t)+J multi (t)

[0128] In the formula: J multi (t) represents polyphonic interference.

[0129] The output signal x(t) after multiplying the above input signal y(t) by the carrier wave and filtering out the high-frequency components by an ideal low-pass filter can be expressed as:

[0130]

[0131] The signal after filtering out high-frequency components is sampled and decided, and the sampling parameter x is in the i-th symbol time interval:

[0132]

[0133] Based on the sign of x, the error symbol probability is calculated, and the bit error rate expression for BPSK under multi-tone interference is obtained as follows:

[0134]

[0135] Among them, T s The modulated symbol width (symbol time width) satisfies T s = (log2M)T b M is the modulation order, T b Δf represents the transmission time width per bit, L represents the number of L interference spectral lines applied, and Δf represents the time width per bit. i J1 represents the frequency interval between the interference signal and the carrier signal, J1 is the power of the multi-tone interference signal, and θ represents the frequency interval between the interference signal and the carrier signal. J1 This represents the phase of the multi-tone interference signal.

[0136] sin c(x) is the sigma function, which satisfies sin c(x) = sin(x) / x.

[0137] (1.3) Theoretical bit error rate of BPSK under narrowband and wideband interference

[0138]

[0139] In the theoretical bit error rate formula involving BPSK:

[0140] P e T represents the bit error rate, A is the amplitude of the BPSK signal, and T is the bit error rate. s Where J is the symbol time width, and J0 is the power of the single-tone interference signal. Let Q be the phase of the single-tone interference signal, and let Q be the right-tail function of the standard normal distribution. Here, L represents the noise variance; L indicates that L interference spectral lines were applied; and J1 is the power of the multitone interference signal. The phase of the multi-tone interference signal, Δf i The frequency interval between the interference signal and the carrier signal is represented by N0; the noise power spectral density is N0; J3 = P J / B J J3 represents the broadband / narrowband interference power spectral density, P J For broadband / narrowband interference signal power, B J For broadband / narrowband interference signal bandwidth, E b This represents the signal energy averaged over each bit.

[0141] [QPSK Disturbance Analysis Model]

[0142] (2.1) Theoretical bit error rate of QPSK under single-tone interference

[0143] The carriers of a QPSK signal are mutually orthogonal, and the rate of the orthogonal input data stream is half that of the baseband data. QPSK signal e QPSK (t) can be represented as:

[0144]

[0145] Where, θ m It is the phase of the QPSK signal, and its value range is... S QPSK It is the power of the signal, w c It is the angular frequency of the signal.

[0146] The signal y(t) at the input of the modulator can be expressed as:

[0147] y(t)=e QPSK (t)+n(t)+J single (t)

[0148] In the formula: J single (t) represents a single-tone interference signal.

[0149] The input signal, after quadrature modulation, is divided into an in-phase branch and a quadrature branch. Both signals are then filtered by ideal low-pass filters to remove high-frequency components. The in-phase branch, after further high-frequency component removal, yields:

[0150]

[0151] After filtering out high-frequency components using orthogonal branches, the following is obtained:

[0152]

[0153] The signal after filtering out high-frequency components is sampled and decided, and the sampling parameter x is in the i-th symbol time interval. I x Q :

[0154]

[0155]

[0156] in: N I N Q The mean is zero and the variance is Gaussian random variable, due to m = 1, 3, 5, 7, therefore a i b i Equal to the in-phase and orthogonal binary sequence, i.e., a i =±1, b i =±1.

[0157] The formula for the bit error rate of a QPSK signal under single-tone interference is:

[0158]

[0159] in:

[0160]

[0161]

[0162] In the formula: N0 is the noise power spectral density.

[0163] (2.2) Theoretical bit error rate of QPSK under multi-tone interference

[0164] The input signal of QPSK under multi-tone interference can be expressed as:

[0165] y(t)=e QPSK (t)+n(t)+J multi (t)

[0166] In the formula: J multi(t) represents the multi-tone interference signal.

[0167] The output signal, after multiplying the input signal y(t) by the carrier wave and filtering out high-frequency components using an ideal low-pass filter, is expressed as:

[0168]

[0169]

[0170] The signal after filtering out high-frequency components is sampled and decided, and the sampling parameter x is in the i-th symbol time interval. I x Q :

[0171]

[0172]

[0173] The formula for the bit error rate of a QPSK signal under multi-tone interference is:

[0174]

[0175] in:

[0176]

[0177]

[0178] (2.3) Theoretical bit error rate of QPSK signal under narrowband and wideband interference

[0179]

[0180] In the formula for the bit error rate of QPSK signals: P e T represents the bit error rate. s Where J is the symbol time width, and J0 is the power of the single-tone interference signal. Let Q be the phase of the single-tone interference signal, and let Q be the right-tail function of the standard normal distribution. Here, L represents the noise variance; L indicates that L interference spectral lines were applied; and J1 is the power of the multitone interference signal. The phase of the multi-tone interference signal, Δf i S represents the frequency interval between the interference signal and the carrier signal. QPSK J3 = P, where N is the power of the QPSK signal and N0 is the noise power spectral density. J / B J J3 represents the broadband / narrowband interference power spectral density, P J For broadband / narrowband interference signal power, B J For broadband / narrowband interference signal bandwidth, E bThis represents the signal energy averaged over each bit.

[0181] [16QAM Disturbance Analysis Model]

[0182] (3.1) Theoretical bit error rate of 16QAM under single-tone interference

[0183] 16QAM signal expression:

[0184]

[0185] Where: δ is the unit amplitude between constellation points, a k =±1,±3,b k =±1,±3,w c Let be the angular frequency of the signal, and t be time. The unit amplitude can be obtained from the 16QAM constellation diagram.

[0186] The input signal at the receiver is: y(t) = s(t) + n(t) + J single (t)

[0187] In-phase branch and carrier Multiplication, orthogonal branches and carrier Multiplying, the outputs of the ideal low-pass filter I and Q paths are respectively:

[0188]

[0189]

[0190] Where: n c (t) represents the in-phase component of the noise, n s (t) represents the orthogonal components of the noise, and J0 represents the power of the single-tone interference. The phase of the single-tone interference signal.

[0191] The signal after filtering out high-frequency components is sampled and decided, and the sampling parameter r is in the i-th symbol time interval. I r Q :

[0192]

[0193]

[0194] Based on the 16QAM constellation diagram, to determine the bit error rate, taking constellation points in the first quadrant as an example, the following four cases are discussed.

[0195] (3.1.1)a k =1, b k =1, the probability of a correct judgment is:

[0196]

[0197] (3.1.2)a k =1, b k =3, the probability of making the correct judgment is:

[0198]

[0199] (3.1.3)a k =3, b k =1, the probability of a correct judgment is:

[0200]

[0201] (3.1.4)a k =3, b k =3, the probability of making the correct judgment is:

[0202]

[0203] in:

[0204]

[0205]

[0206]

[0207]

[0208] The bit error rate of 16QAM under single-tone interference is:

[0209]

[0210] (3.2) Theoretical bit error rate of 16QAM under multi-tone interference

[0211] Under multi-tone interference, the receiver input signal y(t) is:

[0212] y(t)=s(t)+n(t)+J multi (t)

[0213] The output signal, after the input signal is multiplied by the carrier wave and filtered out by an ideal low-pass filter to remove high-frequency components, is expressed as:

[0214]

[0215]

[0216] The signal after filtering out high-frequency components is sampled and decided, and the sampling parameter r is in the i-th symbol time interval. Ir Q :

[0217]

[0218]

[0219] The formula for calculating the bit error rate (BER) of 16QAM under multi-tone interference is the same as that under single-tone interference, where:

[0220]

[0221]

[0222]

[0223]

[0224] (3.3) Theoretical bit error rate of 16QAM signal under narrowband and wideband interference

[0225]

[0226] In the formula for the bit error rate of 16QAM signals: P e T represents the bit error rate. s Where J is the symbol time width, and J0 is the power of the single-tone interference signal. Let Q be the phase of the single-tone interference signal, and let Q be the right-tail function of the standard normal distribution. Here, L represents the noise variance; L indicates that L interference spectral lines were applied; and J1 is the power of the multitone interference signal. The phase of the multi-tone interference signal, Δf i The frequency interval between the interference signal and the carrier signal is represented by N0; the noise power spectral density is S. 16QAM J3 = P, representing the power of a 16QAM signal. J / B J J3 represents the broadband / narrowband interference power spectral density, P J For broadband / narrowband interference signal power, B J For broadband / narrowband interference signal bandwidth, E b This represents the signal energy averaged over each bit.

[0227] Figures 2(a), 2(b), 2(c), and 2(d) show the performance analysis of the BPSK signal after interference with single-tone, multi-tone, wideband, and narrowband signals, respectively. Figures 3(a), 3(b), 3(c), and 3(d) show the performance analysis of the QPSK signal after interference with single-tone, multi-tone, wideband, and narrowband signals, respectively. Figures 4(a), 4(b), 4(c), and 4(d) show the performance analysis of the 16QAM signal after interference with single-tone, multi-tone, wideband, and narrowband signals, respectively. As can be seen from the simulation figures, the simulated bit error rate curves are basically consistent with the derived theoretical curves, proving the rationality of the quantitative analysis of interference on the interfered communication signal in this application. Furthermore, with the increase of the bit signal-to-noise ratio (SNR), the bit error rate under various interference methods gradually decreases; under a fixed SNR, with the increase of the signal-to-interference ratio (SIR), the bit error rate under various interference methods gradually decreases, which is consistent with the actual situation.

[0228] 3. Take effective anti-interference measures

[0229] In communication countermeasures, every detected jamming signal type, signal characteristics, and jamming performance analysis result is invaluable. These results can provide a reference for the selection of communication methods and parameters, and help to implement more effective anti-jamming measures in the future. Therefore, it is necessary to record the signal type, signal characteristics, bit error rate after jamming, and changes before and after jamming for each detected signal, forming a jamming performance analysis model database for reference.

[0230] When our communications are suppressed by the local authorities, we can adjust the communication signals using the recorded interference performance analysis model. This includes changing the communication signal modulation method, signal bandwidth, increasing signal power, and changing the data transmission rate, thereby achieving effective anti-interference.

[0231] [Simulation Experiment]

[0232] (I) The hardware platform for the simulation experiment is: Intel i7-10850H CPU with a main frequency of 2.7GHz and 16GB of memory.

[0233] (II) The software platform is: Windows 10 operating system and Matlab2018b.

[0234] (III) Signal simulation program parameter settings: Signal symbol transmission rate R s =5e6Baud, sampling frequency f s =25e6MHz, carrier frequency f c =5e6MHz, simulated symbol length 1000000, phase difference θ between interference signal and communication signal J=π / 4, base M=2, the signal-to-interference ratio of the system is within 5dB-25dB.

[0235] The deep learning framework PyTorch uses random initialization of weight parameters, cross-entropy as the loss function, a learning rate of 0.001, and a batch size of 64.

[0236] Using MATLAB simulation software, six types of interference signals, including single-tone and multi-tone interference signals, were simulated and generated. In the simulation, the communication signal was set to BPSK, the interference signal sampling frequency was 4MHz, the channel noise was additive white Gaussian noise, and the signal-to-interference ratio (SIR) (the ratio of the power of the communication signal to the power of the interference signal) ranged from -20dB to 10dB with a step size of 2dB. 1500 different samples were generated for each dB increment, with 1000 samples used as the training set and 500 samples used as the test set. The total number of samples in the training set was 96,000, and the total number of samples in the test set was 48,000.

[0237] Depend on Figure 5 The accuracy rates for identifying different types of interference signals show that the accuracy of identifying the six types of interference signals in this experiment increases with the increase of SJR. When SJR is greater than 2dB, the accuracy rates for identifying the six types of interference signals all reach over 90%. This demonstrates that the deep learning-based convolutional neural network interference identification model can effectively extract the features of each interference signal in the spectrogram and has good identification performance. Among them, broadband interference and multi-tone interference have the most obvious features and the best identification results.

[0238] Based on the identification of the type of jamming signal used by the enemy, the anti-jamming performance analysis shows that the bit error rate (BER) of the digital modulation signal is related to the signal-to-noise ratio (SNR), signal-to-interference ratio (SIR), the carrier frequency difference between the jamming and communication signals, the phase difference between the jamming and communication signals, and the baseband data rate. When the carrier frequencies of the jamming and communication signals are the same, and other conditions are constant, the BER changes with the initial phase of the jamming signal. The BER reaches its maximum when the phase difference between the jamming and communication signals reaches π / 4. The stronger the jamming signal, the worse the communication system performance. Therefore, effective anti-jamming can be achieved by changing parameters such as the power, frequency, phase, and symbol transmission rate of the friendly communication signal, depending on the specific scenario.

[0239] Through the above description of the embodiments, those skilled in the art can clearly understand that this disclosure can be implemented using software plus necessary general-purpose hardware, or it can be implemented using dedicated hardware including dedicated integrated circuits, dedicated CPUs, dedicated memory, dedicated components, etc. Generally, any function performed by a computer program can be easily implemented using corresponding hardware, and the specific hardware structure used to implement the same function can be diverse, such as analog circuits, digital circuits, or dedicated circuits. However, for this disclosure, software implementation is more often a preferred implementation method.

Claims

1. A method for quantitative analysis of interference in communication systems for cognitive decision-making, characterized in that, The method includes the following steps: The spectrum of the interference signal is used as input, and the type of interference signal is identified using a pre-defined deep convolutional neural network model. Based on the interfered communication signal, a disturbance analysis model corresponding to the identified interference signal type is used to perform quantitative analysis of the interfered communication signal. If the interfered communication signal is a BPSK signal; When the interference signal type is single-tone interference, the theoretical bit error rate of the BPSK signal is calculated using the following formula: When the interference signal type is multi-tone interference, the theoretical bit error rate of the BPSK signal is calculated using the following formula: When the interference signal type is narrowband interference or wideband interference, the theoretical bit error rate of the BPSK signal is calculated using the following formula: If the interfered communication signal is a QPSK signal; When the interference signal type is single-tone interference, the theoretical bit error rate of the QPSK signal is calculated using the following formula: When the interference signal type is multi-tone interference, the theoretical bit error rate of the QPSK signal is calculated using the following formula: When the interference signal type is narrowband interference or wideband interference, the theoretical bit error rate of the QPSK signal is calculated using the following formula: If the interfered communication signal is a 16QAM signal; When the interference signal type is single-tone interference, the theoretical bit error rate of the 16QAM signal is calculated using the following formula: in: When the interference signal type is multi-tone interference, the theoretical bit error rate of the 16QAM signal is calculated using the following formula: in: When the interference signal type is narrowband interference or wideband interference, the theoretical bit error rate of the 16QAM signal is calculated using the following formula: In the formula: Indicates bit error rate, The amplitude of the BPSK signal. For symbol time width, The power of the single-tone interference signal. For the phase of the single-tone interference signal, The right-tail function of the standard normal distribution. For noise variance; Indicates that it has been applied Root interference spectral lines, The power of the multi-tone interference signal, The phase of the multi-tone interference signal, This indicates the frequency interval between the interference signal and the carrier signal; The noise power spectral density; The power of the QPSK signal; The power of the 16QAM signal; , For broadband / narrowband interference power spectral density, For broadband / narrowband interference signal power, For broadband / narrowband interference signal bandwidth, The signal energy averaged over each bit, This is the Singer function.

2. The method according to claim 1, characterized in that: The disturbance analysis model uses a bandpass filter to filter out noise outside the signal frequency band of the interfered communication signal, modulates it through a modulator, and then filters out high-frequency components of the modulated signal through an ideal low-pass filter. Finally, it samples and decides on the signal after filtering out high-frequency components to obtain a sampling parameter. The error rate of different signals under different interference signals is obtained by calculating the error probability of the symbols based on the symbols of the sampling parameters.

3. The method according to claim 1, characterized in that: After obtaining the theoretical bit error rate of the communication signal after interference, the bit error rate can be reduced by adopting one or more of the following methods in combination: changing the modulation method of the communication signal, the signal power, the signal bandwidth, and the data transmission rate.

4. The method according to claim 1, characterized in that: The types of interference signals include single-tone interference, multi-tone interference, narrowband interference, broadband interference, pulse interference, or targeted interference.

5. The method according to claim 1, characterized in that: The deep convolutional neural network model consists of two convolutional layers, two pooling layers, and three fully connected layers.

6. A computer-readable storage medium, characterized in that: The computer program is stored that can be loaded by a processor and executed according to any one of claims 1 to 5.

Images