Method and system for detecting signal anti-multipath interference based on multimodal feature fusion

Abstract

The present invention relates to a detection signal anti-multipath interference method and system based on multimodal feature fusion, which belongs to the field of radio detection technology and comprises the following steps: step 1: UAV signal fingerprint feature extraction: analyzing the physical layer parameters of the signal; step 2: UAV signal airspace DOA estimation and error analysis: using an array antenna to obtain signal spatial spectrum information; step 3: physical modeling of the ship environment: combining a three-dimensional ship model to construct a multipath reflection path prediction model; step 4: dynamic DOA error correction: suppressing the DOA estimation deviation caused by multipath through iterative optimization of signal features and physical models; step 5: multimodal feature fusion decision: constructing a feature fusion mechanism based on Bayesian probability to distinguish between real signals and false alarm signals. The beneficial effects of the present invention are: achieving the suppression of multipath effects and facilitating UAV signal detection in dynamic and complex scenarios.

Description

Technical Field

[0001] The present invention belongs to the technical field of radio detection, and in particular relates to a method and system for detecting signal anti-multipath interference based on multimodal feature fusion. Background Art

[0002] With the widespread adoption of drone technology, drone signal detection and positioning and direction-finding capabilities have become a core requirement for airspace safety and regulation. Current technology systems are primarily based on traditional radio detection: one relies on physical-layer signal characteristics (such as spectrum energy mutations and time-domain waveform matching) to achieve target identification, while the other relies on geometric positioning principles (such as TDOA time-difference positioning and DOA direction-of-arrival estimation) to achieve spatial positioning and direction-finding. These methods perform well under ideal conditions of line-of-sight propagation and static scenes. However, in highly reflective environments like urban buildings and ships at sea, signal aliasing caused by multipath effects severely limits detection effectiveness. Multipath components generated by electromagnetic waves reflected and diffracted by complex environments not only distort signal characteristics but also create false signal sources, increasing the false alarm rate and degrading positioning accuracy in traditional detection systems.

[0003] There are currently three main problems in suppressing signal multipath effects:

[0004] 1) Signal representation has a single dimension. Traditional methods often analyze time domain, frequency domain, or spatial domain features independently, lacking the ability to jointly model multi-dimensional features and making it difficult to effectively separate direct signals from multipath components.

[0005] 2) Insufficient environmental perception capabilities. The existing static channel assumption does not fully consider the time-varying characteristics of the propagation medium, resulting in poor adaptability to dynamic environments.

[0006] 3) The signal processing dimension is separated. The front-end detection and back-end positioning links usually adopt a cascade processing architecture. The cascade error will accumulate and amplify step by step in the processing link. Summary of the Invention

[0007] The present invention provides a detection signal anti-multipath interference method and system based on multimodal feature fusion, which is used to solve the technical problem of suppressing the multipath transmission effect of drone signals. By establishing a three-dimensional joint analysis framework of "signal fingerprint-spatial wave direction of arrival-environmental physical constraints", the multipath effect is suppressed and drone signal detection in dynamic and complex scenarios is facilitated.

[0008] In order to achieve the above object, the present invention is implemented by the following technical solutions:

[0009] The detection signal anti-multipath interference method based on multimodal feature fusion includes the following steps:

[0010] Step 1: Extracting drone signal fingerprint features: Analyzing the physical layer parameters of the signal;

[0011] Step 2: UAV signal spatial DOA estimation and error analysis: Use array antenna to obtain signal spatial spectrum information;

[0012] Step 3: Physical modeling of the ship environment: Combined with the three-dimensional ship model, a multipath reflection path prediction model is constructed;

[0013] Step 4: Dynamic DOA error correction: Suppress DOA estimation bias caused by multipath through iterative optimization of signal characteristics and physical models.

[0014] Step 5: Multimodal feature fusion decision: Construct a feature fusion mechanism based on Bayesian probability to distinguish real signals from false alarm signals.

[0015] Optionally, in step 1, the physical layer parameter analysis of the signal includes: analysis of the center frequency and bandwidth, extraction of the frequency hopping sequence, and identification of the modulation mode.

[0016] Optionally, the center frequency and bandwidth data are extracted through short-time Fourier transform, and the frequency hopping period and frequency hopping sequence of the signal are detected, and then the modulation mode of the drone signal is identified based on the high-order cumulant.

[0017] Optionally, in step 2, the array antenna is composed of multiple antenna units. When the signal is incident on the array antenna from different directions, different phase delays will be generated due to the different distances between each antenna unit and the signal source. By measuring the phase delay and signal amplitude, an array signal model is established to evaluate the signal source information.

[0018] Optionally, the signal spatial spectrum information includes: constructing an array signal model and multi-signal classification spatial spectrum estimation.

[0019] Optionally, in step 3, the wireless signal is regarded as a light ray, and the propagation path of the wireless signal in the environment is analyzed according to the principle of geometric optics.

[0020] Optionally, in step 3, the multipath reflection path prediction model uses three-dimensional coordinate transformation and multipath reflection path prediction.

[0021] Optionally, in step 4, the DOA estimation deviation caused by multipath is suppressed by jointly estimating multipath parameters and dynamically compensating for the ship platform attitude disturbance.

[0022] Optionally, in step 4, the dynamic DOA error correction adopts characteristic likelihood function modeling and adaptive decision making.

[0023] The detection signal anti-multipath interference system based on multimodal feature fusion includes:

[0024] Extraction module, used for UAV signal feature extraction;

[0025] Analysis module, used for UAV signal airspace DOA estimation and error analysis;

[0026] Physical modeling module, used for physical modeling of ship environment;

[0027] Correction module, used for dynamic DOA error correction;

[0028] Decision module, used for multimodal feature fusion decision;

[0029] The extraction module is connected to the analysis module, the analysis module is connected to the physical modeling module, the physical modeling module is connected to the correction module, and the correction module is connected to the decision module.

[0030] Beneficial effects of the present invention:

[0031] 1. The present invention deeply couples traditionally independently processed signal features with spatial geometric information and dynamic environmental parameters by constructing a multi-dimensional fusion architecture of "signal fingerprint-spatial direction of arrival-environmental physical constraints". This architecture not only uses signal fingerprint features to provide physical layer identity authentication, but also converts multipath reflection path predictions into constraints for DOA estimation through a three-dimensional ship model, enabling the system to actively identify and separate direct signals from multipath interference. Compared with traditional single-dimensional suppression methods, the present invention reduces the false alarm rate while maintaining the detection sensitivity of low signal-to-noise ratio signals.

[0032] 2. The present invention utilizes the combination of dynamic environment coupling mechanism and adaptive decision-making framework to solve the time-varying problem of multipath effect on mobile platforms. By integrating ship attitude data and array manifold correction in real time, it can accurately compensate for the wavefront distortion caused by carrier motion. The multimodal decision-making strategy can dynamically adjust the classification threshold according to the signal-to-noise ratio and environmental stability parameters. This closed-loop optimization mechanism enables the system to maintain stable signal detection performance in complex electromagnetic environments and dynamic motion scenarios, significantly improving the applicability of existing UAV detection systems for ship platforms in strong reflection scenarios.

[0033] 3. The modular technical architecture proposed in this invention combines engineering practicality with technical scalability. The modules in the signal processing chain adopt a loosely coupled design. It can be quickly deployed based on the existing TDOA / DOA hardware platform, and is compatible with the expanded access of new signal characteristics in the future. This makes the system widely applicable in the fields of port security and urban low-altitude supervision, and provides key technical support for building an all-weather, all-airspace drone monitoring network. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0035] Figure 1 It is a system structure diagram of the present invention;

[0036] Figure 2 It is the workflow diagram of the present invention. DETAILED DESCRIPTION

[0037] The embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0038] Multipath effect: Electromagnetic waves propagate through different paths (such as reflection and diffraction) before reaching the receiving end, resulting in signal delay expansion, phase distortion, and energy superposition. The interference between multipath components and the direct signal can cause amplitude fluctuations in the received signal, waveform distortion, and spatial positioning deviation.

[0039] Signal fingerprint feature extraction: By extracting the inherent characteristics of the signal physical layer (such as the spectrum envelope, modulation parameters or frequency hopping sequence), a unique identification technology is established to achieve transmitter identity authentication and interference signal identification.

[0040] Frequency hopping sequence extraction: Extracts spread spectrum communication signals whose carrier frequency periodically jumps according to a predetermined random sequence. The instantaneous bandwidth of the spread spectrum communication signal varies, but the total bandwidth remains unchanged. Frequency hopping can avoid interference and enhance communication anti-interception capabilities.

[0041] Support Vector Machine (SVM) Classifier: A supervised learning model based on structural risk minimization. It uses kernel functions to map low-dimensional inseparable data to a high-dimensional feature space, constructing an optimal hyperplane to achieve pattern classification. In pattern classification and recognition, it uses the principle of maximizing classification intervals to improve generalization capabilities under small sample conditions.

[0042] MUSIC (Multiple Signal Classification): This algorithm combines high-resolution spatial direction of arrival (DOA) with a spatial DOA estimation algorithm based on subspace decomposition. By analyzing the covariance matrix of the received signal, it separates the signal and noise subspaces and constructs a spatial spectrum function using the orthogonality of the signal and noise subspaces.

[0043] Example 1;

[0044] like Figure 1 As shown in FIG, the detection signal anti-multipath interference system based on multimodal feature fusion includes:

[0045] Extraction module, used for UAV signal feature extraction;

[0046] Analysis module, used for UAV signal airspace DOA estimation and error analysis;

[0047] Physical modeling module, used for physical modeling of ship environment;

[0048] Correction module, used for dynamic DOA error correction;

[0049] Decision module, used for multimodal feature fusion decision;

[0050] The extraction module is connected to the analysis module, the analysis module is connected to the physical modeling module, the physical modeling module is connected to the correction module, and the correction module is connected to the decision module.

[0051] The work is carried out in the order of extraction module, analysis module, physical modeling module, correction module and decision module.

[0052] Example 2;

[0053] Based on Example 1, Figure 2 As shown, this embodiment provides a detection signal anti-multipath interference method based on multimodal feature fusion, including the following steps:

[0054] Step 1: Extracting drone signal fingerprint features: Analyzing the signal's physical layer parameters. This includes center frequency and bandwidth analysis, frequency hopping sequence extraction, and modulation mode identification.

[0055] Drone signal feature extraction:

[0056] The center frequency and bandwidth data are extracted through short-time Fourier transform (STFT), and the frequency hopping period and frequency hopping sequence of the signal are detected. Then, the modulation mode of the drone signal is identified based on the high-order cumulant (HOC).

[0057] (1) Center frequency and bandwidth analysis;

[0058] Assume that the discrete time signal sequence received by the radio detection equipment is , short-time Fourier transform calculation formula:

[0059] ;

[0060] in, is a discrete time index, representing the sampling points of the signal at different times; is a continuous time variable, that is, at different moments, and is also the center position of the window function. The window function can be moved on the time axis to perform spectrum analysis on different time periods of the signal. Usually rounded to an integer; is the window function, and the discrete time signal sequence is There is overlap, and the step size of each sliding of the window function can be adjusted as needed. If the step size is 1, the time resolution of the analysis is the highest, but the amount of calculation is also the largest; if the step size is greater than 1, the amount of calculation will be reduced, but the time resolution will be reduced;

[0061] is a complex exponential function, is an imaginary unit used for spectrum analysis. is the number of points in the discrete Fourier transform (DFT).

[0062] Specifically, the Hanning window can reduce spectrum leakage. Each frame of the signal is weighted using the Hanning window to balance the main lobe width and sidelobe suppression capability. The window size can be dynamically adjusted according to the frequency hopping speed. The typical value is 5ms, which is suitable for drone signal analysis with a frequency hopping period ≥10ms.

[0063] is the frequency index, by calculating the spectrum amplitude , a time-frequency diagram can be generated. In the time-frequency diagram, for each time slice (corresponding to a frame signal, a frame signal is ) Perform spectrum peak detection, retain the frequency points whose amplitude exceeds the detection threshold as candidate frequency modulation points, and obtain the center frequency point of the signal through spectrum peak detection and bandwidth .

[0064] Center frequency Refers to the center position of each frequency interval. For the STFT of discrete-time signals, the frequency index The corresponding center frequency The formula can be It is calculated that, A frame signal is The sampling frequency, is the number of points at which the Discrete Fourier Transform (DFT) is performed.

[0065] bandwidth Refers to the frequency range covered by each frequency interval. For the STFT of the window function, its bandwidth is related to the spectrum characteristics of the window function and the number of DFT points. Find the minimum frequency point among the candidate frequency modulation points and maximum frequency , then the signal bandwidth It can be expressed as: .

[0066] The center frequency determines the position of the bandwidth in the entire spectrum. The choice of center frequency and bandwidth will directly affect the performance of the communication system. A higher center frequency can provide a wider available bandwidth.

[0067] (2) Frequency hopping sequence extraction;

[0068] If there are multiple candidate frequency hopping points in a single frame, the point with the highest amplitude is selected as the current frequency hopping frequency. In order to eliminate false frequency hopping caused by transient interference or noise, the current frequency hopping frequency can be combined with adjacent candidate frequency hopping points and smoothed through a sliding window.

[0069] Frequency hopping period estimation:

[0070] Statistical frequency hopping switching time , then calculate the time interval sequence Autocorrelation function, find the period value corresponding to the periodic peak , and count the majority as the frequency hopping period :

[0071] ;

[0072] in, It is a statistical distribution function that is mainly used to return the most frequently occurring value in a data set, that is, the mode.

[0073] Frequency-Channel Mapping :

[0074] ;

[0075] in, is the lower frequency limit, is the channel spacing, is the frequency occupied by each channel, Indicates rounding down.

[0076] (3) Signal modulation identification;

[0077] The identification of the modulation mode of the drone signal is achieved based on the high-order cumulant (HOC), including the identification of the modulation types of BPSK, QPSK and OFDM.

[0078] Intercept signal segment (i.e. discrete-time signal sequence), calculate the fourth-order cumulant and :

[0079] ;

[0080] ;

[0081] in, is the joint cumulant function, for The complex conjugate of for The fourth-order origin moment of for The second-order central moment variance of for The second-order origin moment variance of .

[0082] In practical applications, finite sample data are usually selected to estimate the fourth-order cumulant and :

[0083] ,in, for The fourth-order origin of for The second-order center of is the total number of finite sample data;

[0084] ,in, for The second-order origin of

[0085] For common signal modulation methods, the fourth-order cumulant value is:

[0086] QPSK modulation: ;

[0087] BPSK modulation: ;

[0088] 16-QAM modulation: ;

[0089] OFDM modulation: .

[0090] According to the high-order cumulative amount characteristics of different modulation modes, the modulation mode can be distinguished by using the support vector machine (SVM) classifier.

[0091] Step 2: UAV signal spatial DOA estimation and error analysis: Use array antennas to obtain signal spatial spectrum information. This includes constructing an array signal model and multi-signal classification spatial spectrum estimation.

[0092] An antenna array consists of multiple antenna elements. When signals strike the array from different directions, different phase delays are generated due to the varying distances between each antenna element and the signal source. By measuring these phase delays and signal amplitudes, a model of the array signal can be constructed, allowing the source information to be assessed.

[0093] UAV signal airspace DOA estimation and error analysis:

[0094] (1) Construct array signal model;

[0095] Assume that the radio signal receiving device receives signal source, the array output signal is:

[0096] ;

[0097] in, For continuous time The signal received at any time, is the array manifold matrix, that is, From the direction The signal response, From the direction The signal response, is a source signal in the far field, is the additive noise on the array element, for time, , It is The direction of arrival of the signal, usually the angle relative to the array normal, is the wavelength of the signal, is the spacing between adjacent array elements, is an imaginary unit, , To describe the The geometric relationship between the arrival direction of the signal and the array is calculated, that is, the path difference between each array element caused by the different signal incident angles is calculated, and then the phase difference is obtained. is the conjugate transpose.

[0098] (2) MUSIC (Multiple Signal Classification) spatial spectrum estimation;

[0099] The covariance matrix of the received signal is calculated as:

[0100] ;

[0101] Among them, the covariance matrix Is a symmetric matrix. In physical terms, the signal and The covariance and and The covariances of are the same, reflecting the correlation between the two signal elements, so the covariance matrix is symmetric about the main diagonal. For signal, is the conjugate transpose of the signal; represents the mathematical expectation, It's a signal The mean vector of ;

[0102] is the covariance matrix For any non-zero vector , both ≥0, this property can be understood from the definition of covariance, As a signal After linear transformation A measure of the variance after , the variance is always non-negative, so a≥0. From the perspective of the matrix, semi-positive definiteness means that the eigenvalues of the covariance matrix are all greater than or equal to zero, which ensures the stability and solvability of the covariance matrix in mathematical operations and signal processing applications.

[0103] Diagonal elements of the covariance matrix It's a signal Variance , which indicates the signal The degree of fluctuation around its mean reflects the signal The larger the variance of the energy distribution in a certain dimension, the more drastic the signal change in that dimension, and the more energy it carries. for Received signal The number of samples.

[0104] By decomposing the eigenvalue, the noise subspace is obtained and the spatial spectrum function is constructed as follows:

[0105] ;

[0106] in, Reflected in different angles The distribution of signal energy on the From a certain angle When incident on the array, if the angle matches the actual arrival angle of the signal source, then the spatial spectrum function will have a peak at this angle.

[0107] Denominator of the spatial spectrum function Represents an array manifold vector In the noise subspace matrix The square of the projection length on is the array manifold vector The conjugate transpose of is the noise subspace matrix The conjugate transpose of is the vector projection.

[0108] When it is 0, tends to infinity, forming a peak, and when When it is not the angle of arrival of the signal source, is not completely orthogonal to the noise subspace, When greater than 0, is a finite value and is relatively small. Therefore, by searching The peak position of the signal can be used to estimate the arrival angle of the signal source.

[0109] The spatial spectrum function divides the observation space into signal subspace and noise subspace based on eigenvalue decomposition, and uses the projection characteristics of the array manifold vector on the noise subspace to estimate the signal source arrival angle through the peak value of the spatial spectrum function.

[0110] (3) DOA error caused by multipath effect;

[0111] When there is When there are multiple multipath components, the actual array flow deformation for:

[0112] ;

[0113] in, 、 Respectively The attenuation coefficient and phase offset of each path, is the angle of incidence corresponding to The single-path steering vector, is an imaginary unit. The traditional MUSIC algorithm will produce a pseudo peak under this condition, resulting in a large deviation in the angle estimation. Representative The phase factor of the multipath component, is the phase offset of the multipath signal relative to the reference phase;

[0114] The actual received array manifold vector is The result of the interaction of the multipath components is the steering vector corresponding to each multipath component. Amplitude-weighted and phase adjustment Each multipath component contributes to the total array manifold, and its contribution is determined by amplitude and phase.

[0115] Step 3: Physical modeling of the ship environment: Combined with the three-dimensional ship model (including real-time attitude data of three-dimensional coordinate transformation), a multipath reflection path prediction model is constructed; that is, the wireless signal is regarded as a light ray, and the propagation path of the signal in the environment is analyzed based on the principles of geometric optics, such as reflection, refraction, and diffraction.

[0116] Physical modeling of the ship environment:

[0117] (1) Three-dimensional coordinate transformation;

[0118] Establish the ship's coordinate system With the global coordinate system The conversion relationship:

[0119] ;

[0120] Among them, the rotation matrix of the ship's global coordinate system is , the rotation matrix of the ship body coordinate system is , is the attitude rotation matrix, i.e. real-time attitude data, For rolling, For pitch, For the bow, is the yaw angle around the z axis, is the pitch angle around the y-axis, is the roll angle around the x-axis, is the translation vector.

[0121] As a ship sails at sea, its position and attitude constantly change. Using conversion formulas, motion parameters in the ship's body coordinate system, such as velocity, acceleration, and attitude angle, can be converted to a global coordinate system. This allows for an accurate description of the ship's motion within a unified global reference frame, facilitating navigation, monitoring, and command.

[0122] (2) Multipath reflection path prediction;

[0123] Based on the theory of geometric optics and Fermat's principle, the ray tracing method is used to solve the reflection path of the radio signal through the rigid structure of the ship.

[0124] Solving the signal reflection path requires satisfying two conditions:

[0125] Path length extreme condition: reflection point Make the total transmission path Get the minimum value, is the launch point, For the receiving point; for Point to the magnitude (i.e., length) of the point's path vector, for Point to The magnitude of the point's path vector.

[0126] The physical meaning of this extremum condition is that when light (or a radio signal) propagates along a path that satisfies Fermat's principle, the path length reaches an extremum. Fermat's principle states that when light propagates through a medium, the actual path is the one that causes the optical path length to reach an extremum (a minimum, maximum, or stable value). In reflection problems, this means that the path of the reflected light is such that the total path length satisfies certain extremum conditions, thereby determining the location of the reflection point.

[0127] Snell's law of reflection: the angle of reflection is equal to the angle of incidence, that is, the reflection point The surface normal vector at satisfy:

[0128] ;

[0129] in, for Point to The path of the point, for Point to The path of the point;

[0130] 、 are all normalized unit vectors, that is, is the incident unit vector, is the reflected unit vector, satisfying 、 Both are 1.

[0131] After normalization, the vector only retains the direction information and the length is unified to 1. After normalization, the normal vector can be used for dot product operations in lighting calculations (such as calculating the cosine values of the incident angle and the reflection angle).

[0132] The reflection path optimization problem is defined as:

[0133] ;

[0134] in, is the optimal reflection point, is the reflection point, is the surface point set of the ship 3D model, for Point to The square of the magnitude of the point's path vector, for Point to The square of the magnitude of the point's path vector; To calculate the optimal reflection point.

[0135] By introducing the Lagrangian operator (Lagrange multipliers), with the law of reflection as a constraint :

[0136] ;

[0137] in, is the incident direction vector, is the reflection direction vector, is a nonlinear constraint, To minimize the optical path length (i.e., minimize the path length).

[0138] The Lagrange multiplier method takes the law of reflection as a constraint. By constructing the Lagrange function and solving the extreme value conditions, a path that satisfies the law of reflection can be found.

[0139] Reflection point The stationary point condition is obtained by taking the derivative at:

[0140] ;

[0141] in, is the optical path, which is the sum of the incident unit vector and the reflected unit vector, is the normal component, that is component in the opposite direction of the incident unit vector, and The component is in the opposite direction of the reflected unit vector.

[0142] Indicates at the reflection point At this point, the incident unit vector is parallel to the component in the direction opposite to the incident unit vector, and the reflected unit vector is parallel to the component in the direction opposite to the reflected unit vector, which means that the stationary point condition automatically implies the law of reflection.

[0143] Simplifying the vector form of the reflection law:

[0144] ;

[0145] The left side of this formula is the sum of the incident unit vector and the reflected unit vector, and the right side is the normal vector The scaling factor is .

[0146] This equation accurately describes the geometric relationship at the reflection point, and is calculated by numerical calculation (such as Newton iteration method) on the surface point set of the three-dimensional ship model. The above solution obtains the effective reflection point set.

[0147] Step 4: Dynamic DOA error correction: Suppress DOA estimation bias caused by multipath through iterative optimization of signal characteristics and physical models.

[0148] Dynamic DOA error correction:

[0149] (1) Joint estimation of multipath parameters;

[0150] In the radio detection array antenna receiving signal model, it is assumed that there is multipath transmission components, the output signal of the receiving antenna array is expanded to:

[0151] ;

[0152] in, For continuous time The signal received at any time, When it is 0, it corresponds to the direct path of the signal. It is The amplitude coefficient of a signal component controls the intensity of the corresponding signal component and is a real number. is the complex attenuation coefficient, is a vector, called the array manifold vector, It is with Parameters related to the signal components, such as the direction of arrival (DOA) of the signal in array signal processing, the vector It reflects the spatial characteristics of the signal on the array.

[0153] is a complex exponential term, is an imaginary unit, is the phase offset, that is It is The phase of a signal component. This complex exponential term is used to represent the phase information of the signal. In signal processing, the phase has an important influence on the synthesis and interference characteristics of the signal.

[0154] is the source signal (or emission signal), which is a signal about the time function, which describes the change of the signal over time.

[0155] It is additive noise, and it is also about the moment In the actual signal reception process, noise is an inevitable interference factor, which will be superimposed on the useful signal and affect the detection and processing of the signal.

[0156] Set up a parameter estimation optimization problem:

[0157] ;

[0158] in, For signal, this formula is an optimization method, by selecting a suitable set of parameters , to minimize an expression, where min means to find the minimum value, Represents the norm of the vector (Euclidean norm), which is used to measure and The "distance" between them is the difference between them. The goal of optimization is to minimize this difference. Is a summation operation, which means Multipath transmission components Add up, by adjusting 、 and , a desired signal form can be synthesized.

[0159] Limited by the physical constraints of the ship platform environment:

[0160] ;

[0161] Represents a collection, It is a and is a set of parameter definitions, and the constraint set is composed of the surface point set of the ship 3D model With real-time pose matrix Jointly determine, is a set of symbols, belong , by parameterizing the feasible solution space to exclude physically unreachable reflection paths, we can express signal components The surface point set of the 3D model must be With real-time pose matrix range.

[0162] (2) Dynamic compensation of ship platform attitude disturbance;

[0163] The change of the ship platform's attitude will cause the array antenna receiving signal manifold to produce time-varying distortion, and the array position disturbance model in the body coordinate system is defined. :

[0164] ;

[0165] in, is the Jacobian matrix of the ship platform attitude angular velocity, is the Jacobian matrix, is the attitude angular velocity vector of the ship platform, Representation matrix It is a real matrix with 3 rows and 3 columns. In the ship attitude problem, it usually corresponds to the conversion relationship between the angular velocity related to the three attitude angles of roll, pitch, and yaw;

[0166] is the attitude angle disturbance, is the roll angle, is the pitch angle, is the yaw angle, The attitude angle disturbance is the transpose of a vector. It refers to the small change in the attitude angle (such as pitch angle, yaw angle and roll angle) of an object in space relative to a reference attitude. It is used to accurately describe the degree of deviation of the object's attitude from the ideal or nominal attitude.

[0167] The array manifold correction term is:

[0168] ;

[0169] Among them, the array manifold correction term Used to describe the direction of the array antenna The response error or disturbance on Indicates that due to the position error of the array element The resulting array manifold (i.e., the array in the direction The correction amount for each element is the ideal response on Position deviation A phase error will be introduced, and the correction term is obtained by superimposing the contributions of all array elements.

[0170] is the array element position error In the direction The linear phase perturbation on It's a formation element In an ideal location The original phase delay at is an imaginary unit, It is for all The contributions of each array element are summed to obtain the total correction.

[0171] is the unit vector of the signal beam arrival direction, is the transpose of the vector, is the wavelength, and wavelength Combined with, it represents the direction of signal propagation in space, is the projection component of the incident direction on the y-axis, is the projection component of the incident direction on the x-axis.

[0172] Array manifold after real-time compensation for:

[0173] ;

[0174] The error term is Superimposed on the ideal manifold, dynamic correction of array response is achieved, thereby improving the robustness and accuracy of algorithms such as beamforming and direction finding;

[0175] Indicates the time-varying or fixed error introduced by factors such as hardware defects, temperature changes, and signal interference in the actual environment;

[0176] Indicates that under the condition of no error, the direction of the array antenna is The response is usually a complex vector containing the amplitude and phase information of each array element;

[0177] By estimating the attitude disturbance through an online Kalman filter, dynamic calibration of the array response is achieved. Through real-time compensation, the system can more accurately reflect the response characteristics of the actual array and improve performance in complex environments.

[0178] Step 5: Multimodal feature fusion decision: Construct a feature fusion mechanism based on Bayesian probability to distinguish real signals from false alarm signals.

[0179] Multimodal feature fusion decision:

[0180] (1) Feature likelihood function modeling;

[0181] Define the hypothesis space as:

[0182] : The observation signal comes from a real drone;

[0183] : The observed signal is a multipath false alarm;

[0184] Based on signal spectrum characteristics , modulation characteristics Based on the conditional independence assumption of the DOA estimator, a joint feature likelihood function model is established:

[0185] ;

[0186] in, for categories, To estimate the parameters, and Characteristics and features In the assumption Likelihood under;

[0187] For parameter estimation exist The distribution under , reflecting the uncertainty or prior knowledge of the parameters.

[0188] Indicates that Next, Features 、 and estimated parameters The joint probability distribution of .

[0189] Signal spectrum feature likelihood function:

[0190] ;

[0191] in, Indicates that Signal spectrum characteristics The mean value of When it is established, the average characteristics of the signal spectrum characteristics, It is assumed The signal spectrum characteristics The covariance matrix of The inverse matrix of The signal spectrum characteristics are calculated and mean half the square of the Mahalanobis distance, Measures the distance of a point to a distribution, Indicates a proportional relationship.

[0192] This formula is based on the probability density function form of the multivariate normal distribution. In signal processing, it is often assumed that the signal spectrum characteristics obey the normal distribution, and the likelihood function is used to measure the probability of the signal being normal. When the signal spectrum characteristics are observed The size of the possibility.

[0193] Signal modulation feature likelihood function:

[0194] ;

[0195] in, Is the multiplication symbol, indicating the arrive The terms of the multiplication operation means that the characteristics Multiple sub-features Here, the contribution of each sub-feature to the likelihood probability is comprehensively considered through multiplication.

[0196] is the output of the support vector machine (SVM). SVM is a machine learning classification algorithm that classifies sub-features As input, assuming Under the process, a numerical value related to the sub-feature under a specific assumption is output, which can be understood as the sub-feature In the assumption A certain metric value belonging to a certain category (corresponding to a specific modulation method).

[0197] is the weight coefficient used to measure each sub-feature The relative importance of the probability to the overall likelihood calculation. The value can be obtained based on prior knowledge or training to highlight the role of certain more discriminative features.

[0198] DOA space estimation likelihood function:

[0199] ;

[0200] in, is the normalization constant of the Gaussian distribution, which ensures that the integral of the entire probability density function on the domain is equal to 1, so that the probability has reasonable statistical significance.

[0201] is the predicted direction of arrival parameter value, assuming Bottom, the direction of arrival predicted by the model;

[0202] is the estimated parameter The variance measures the degree of dispersion of the estimated value. The smaller the variance, the more concentrated the estimated parameters are. The larger the variance, the more dispersed the distribution of the estimated parameters. The impact of the degree of deviation from on the probability. The smaller the deviation, the closer the value is to 1, and the greater the probability density.

[0203] It is the probability density function of the Gaussian distribution, which reflects the influence of the degree of deviation of the data from the mean on the probability density. The greater the deviation, the smaller the value of the estimated parameter and the corresponding probability density.

[0204] This formula is used to describe the assumption Next, estimate the parameters The likelihood function measures the probability of observing the current estimated value under a given model and parameter assumptions. By maximizing the likelihood function, the most likely direction of arrival estimate can be found.

[0205] (2) Adaptive decision-making;

[0206] Construct the dynamic threshold function as:

[0207] ;

[0208] in: is the dynamic threshold, It is an initial, fixed parameter, usually a constant, which acts as a baseline scaling factor and determines the approximate magnitude of the entire function range.

[0209] : Time-varying noise power estimation, expressed at time The power of the noise;

[0210] : Signal power moving average, representing the time The power sliding average of the signal;

[0211] It is the ratio of time-varying noise power to signal power. When the time-varying noise power is larger than the signal power, it means that the signal is seriously interfered by noise. In this case, the threshold needs to be increased to reduce the impact of noise on subsequent processing.

[0212] When the signal power is larger than the time-varying noise power, it indicates that the signal quality is good and the threshold can be appropriately lowered to detect signal changes more sensitively.

[0213] : The norm of the ship platform attitude change rate, which measures the The rate of change in space (if spatial dimensions are involved) or other relevant dimensions, is the gradient, is the real-time posture matrix;

[0214] Denominator along with , then The overall decrease will be When the rate of change is large, the signal can be detected more sensitively in areas with drastic changes.

[0215] when When it is larger, it indicates The changes are drastic. The value of is small, the dynamic threshold The signal intensity will be reduced accordingly, so that the signal changes can be detected more sensitively in areas with drastic changes, avoiding missing important information.

[0216] when When smaller, The change is slow, The value of is large, the dynamic threshold It will be increased accordingly to reduce false detection in stable areas.

[0217] The final decision rule is defined as:

[0218] ;

[0219] For a given feature 、 and estimated parameters Under the condition of The probability of establishment, For a given feature 、 and estimated parameters Under the condition of The probability of establishment.

[0220] It is a conceptual expression of the relative status of two hypotheses, helping to measure the tendency to support a hypothesis;

[0221] To determine the operation, The value is greater than the threshold When the condition If The value is less than the threshold When the condition is established, that is, the observed signal is a multipath false alarm.

[0222] Can be a moment A function that affects the trade-off between the probability ratios of two hypotheses.

[0223] Dynamically adjust the dynamic threshold by online estimation of ship platform environmental parameters , while ensuring the line of sight of detection probability, effectively control the false alarm rate caused by the signal multipath transmission effect.

[0224] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope of the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A detection signal anti-multipath interference method based on multimodal feature fusion, characterized in that: The steps include: Step 1: Extracting drone signal fingerprint features: Analyzing the physical layer parameters of the signal; Step 2: UAV signal spatial DOA estimation and error analysis: Use array antenna to obtain signal spatial spectrum information; Step 3: Physical modeling of the ship environment: Combined with the three-dimensional ship model, a multipath reflection path prediction model is constructed; Step 4: Dynamic DOA error correction: Suppress DOA estimation bias caused by multipath through iterative optimization of signal characteristics and physical models. Step 5: Multimodal feature fusion decision: Construct a feature fusion mechanism based on Bayesian probability to distinguish real signals from false alarm signals.

2. The detection signal anti-multipath interference method based on multimodal feature fusion according to claim 1 is characterized in that: In step 1, the physical layer parameter analysis of the signal includes: center frequency and bandwidth analysis, frequency hopping sequence extraction, and modulation mode identification.

3. The detection signal anti-multipath interference method based on multimodal feature fusion according to claim 2 is characterized in that: The data of the center frequency and the bandwidth are extracted by short-time Fourier transform, and the frequency hopping period and the frequency hopping sequence of the signal are detected. Then, the modulation mode of the drone signal is identified based on the high-order cumulant.

4. The detection signal anti-multipath interference method based on multimodal feature fusion according to claim 1 is characterized in that: In step 2, the array antenna is composed of multiple antenna units. When the signal is incident on the array antenna from different directions, different phase delays will be generated due to the different distances between each antenna unit and the signal source. By measuring the phase delay and signal amplitude, an array signal model is established to evaluate the signal source information.

5. The detection signal anti-multipath interference method based on multimodal feature fusion according to claim 4 is characterized in that: The signal spatial spectrum information includes: constructing an array signal model and multi-signal classification spatial spectrum estimation.

6. The method for detecting signals against multipath interference based on multimodal feature fusion according to claim 1, wherein: In step 3, the wireless signal is regarded as a light ray, and the propagation path of the wireless signal in the environment is analyzed according to the principle of geometric optics.

7. The method for detecting signals against multipath interference based on multimodal feature fusion according to claim 6, characterized in that: In step 3, the multipath reflection path prediction model uses three-dimensional coordinate transformation and multipath reflection path prediction.

8. The method for detecting signals against multipath interference based on multimodal feature fusion according to claim 1, wherein: In step 4, the DOA estimation deviation caused by multipath is suppressed by jointly estimating multipath parameters and dynamically compensating for the ship platform attitude disturbance.

9. The method for detecting signals against multipath interference based on multimodal feature fusion according to claim 1, wherein: In step 4, the dynamic DOA error correction adopts characteristic likelihood function modeling and adaptive decision making.

10. A detection signal anti-multipath interference system based on multimodal feature fusion, used to execute the detection signal anti-multipath interference method based on multimodal feature fusion according to any one of claims 1 to 9, characterized in that: include: Extraction module, used for UAV signal feature extraction; Analysis module, used for UAV signal airspace DOA estimation and error analysis; Physical modeling module, used for physical modeling of ship environment; Correction module, used for dynamic DOA error correction; Decision module, used for multimodal feature fusion decision; The extraction module is connected to the analysis module, the analysis module is connected to the physical modeling module, the physical modeling module is connected to the correction module, and the correction module is connected to the decision module.

Images