105 results about "Orthogonalization" patented technology

The invention relates to the technical field of data processing, in particular to a backtracking analysis model construction method based on an attack chain, which comprises the following steps that: a kernel layer security agent acquires process, file and network behavior characteristics in a hardware isolation environment, and generates an event tuple; the tensor network pipeline performs three-dimensional decoupling mapping on the tuple into a behavior fingerprint vector, an orthogonalization noise feature and an asymmetric adjacent tensor, and compresses the behavior fingerprint vector, the orthogonalization noise feature and the asymmetric adjacent tensor into a space-time topology tensor block; the reinforcement learning controller constructs a directed acyclic graph based on the tensor blocks, calculates connectivity loss and outputs an event risk score; the dynamic routing engine constructs a decision tree model according to the risk mark, the burst frequency and the correlation entropy, and implements three-level shunting and a multiple simulation system to generate an anti-interference index; and when the deviation between the physical trajectory and the digital model exceeds the tolerance, the closed-loop feedback weight coefficient updates the loss function parameter and adjusts the channel resource weight. And the problem of threat discovery delay caused by attack chain breakage under massive events is solved.

The invention relates to the technical field of electrical digital data processing, and discloses a source code security vulnerability semantic detection method based on a large language model, which comprises the following steps: analyzing a source code to be detected to extract execution path constraint features, and constructing an orthogonal feature base vector sequence through orthogonalization feature extraction; inputting the source code to be tested into the large language model to obtain an initial semantic tensor; orthogonal projection is carried out on the initial semantic tensor to a code security constraint subspace constructed by an orthogonal feature basis vector sequence, weighted aggregation is carried out in combination with attention weight distribution information entropy, and a refined semantic vector is generated; according to the hidden logic offset vulnerability recognition method, business semantic noise is eliminated by utilizing a logic subspace projection mechanism, the association between a detection conclusion and code execution logic is established, and the precision of recognizing hidden logic offset vulnerabilities is improved.

The invention discloses a seal removing and document repairing method based on quantum state collaborative regulation and control, and relates to the field of document image processing and quantum computing cross technology, the method comprises the following steps: obtaining to-be-processed information, and carrying out quantum-classical feature collaborative preparation; the character stroke continuity is guaranteed through quantum entangled state modeling and entanglement degree constraint iteration, texture decoupling is achieved through quantum wavelet transform and Gram-Schmidt orthogonalization, and adaptive filling is conducted in combination with a quantum generative adversarial network; dynamic quantum phase adjustment is used for counteracting superposition interference of stamps with different transparency, and quantum neural network noise reduction and multi-scale quantum Fourier sharpening are used for optimizing image quality; checking the repair result, if the repair result does not reach the standard, returning to the edge sharpening link to perform decoupling and filling the edge sharpening link to readjust the parameter; and for special scenes such as inclination, multi-color overprinting and ultra-thin frames, quantum rotation correction, color channel separation and boundary annihilation operator processing are used, finally, high-precision, high-naturalness and high-adaptability restoration of seal removal is achieved, and high fidelity of results is guaranteed.

The invention discloses an interpretable load prediction method based on knowledge distillation and a time self-attention additive neural network, and the method comprises the steps: collecting the historical load and meteorological data of a power grid as the input characteristics of a model, carrying out the detection of a data quartile abnormal value, dividing the data quartile abnormal value into a training set, a test set and a verification set, and carrying out the detection of the data quartile abnormal value; standardization and abnormal value filling are carried out through Z-shaped orthogonalization and linear filling, and finally, a tensor form meeting the model input requirement is converted through a sliding window; designing a knowledge distillation'teacher-student 'framework based on multiple scales and multiple cycles; constructing a time self-attention additive neural network TSA-NAM as a student model; calculating a shape function representing the contribution degree and the characteristic value in the sub-network to obtain the interpretability of the characteristic dimension; exporting the attention weight of the time self-attention module to obtain the interpretability of the time dimension; performing simulation verification; according to the method, high reliability and high precision are guaranteed, and meanwhile, multi-dimensional interpretability is brought to power load prediction.

The invention discloses a flood storage and detention area change pattern spot identification label generation technology based on style decoupling. The technology comprises the following steps: S1, constructing and preprocessing a change detection data set; s2, constructing a conditional diffusion generation network based on a decoupling encoder; s3, decoupling extraction and orthogonalization representation of content-style features are carried out; s4, constructing a multi-target training strategy and two-stage model training; s5, injecting and fusing style features based on cross attention; s6, cross-domain style migration and diversified label generation; and S7, based on label quality screening of physical and semantic double constraints, outputting a high-quality change detection expansion data set. Compared with the prior art, the method has the advantages that by introducing a content-style decoupling mechanism, the style and the content of the generated sample are independently and accurately controlled, and the change detection label which is consistent in ground feature layout, diversified in imaging style and accurately labeled at a pixel level is generated.

The invention provides an M-system chaos shift keying communication method, device, equipment and medium, and the method comprises the steps: generating a long chaos sequence and a short chaos sequence through employing two chaos generators with different chaos mapping, mapping an M-system symbol into an embedded position through employing a symbol-to-position mapping table, and carrying out the mapping of the M-system symbol into the embedded position; a short chaotic sequence is embedded into a corresponding position of a long chaotic sequence to generate a combined chaotic transmitting signal, so that M-system transmission can be realized only through two chaotic generators, and M chaotic mapping is not needed to construct a signal space; and a receiving end adopts a deep neural network demodulation model to carry out probability judgment on the segmented subsequences, so that incoherent demodulation is realized, the traditional chaotic synchronization technology and orthogonalization technology are not needed, the system structure is effectively simplified, and the realization difficulty is reduced.

The invention discloses a coal mine production anomaly detection system and method based on time series decomposition and prediction, and provides an end-to-end decomposition-prediction anomaly detection framework, the framework firstly constructs a deep STL iterative decomposition network, and uses a differentiable median filtering and iterative orthogonalization mechanism to obtain a deep STL iterative decomposition network; an original power load sequence is decoupled into three orthogonal components, namely a trend component, a season component and a residual error in a robust manner, and modal aliasing is effectively inhibited; then, a parallel multi-branch prediction architecture is adopted, aiming at the physical characteristics of each component, an implicit neural predictor based on a polynomial basis / Fourier basis and a hierarchical frequency domain Transform-CNN are respectively used for special modeling and prediction, and anomaly judgment is realized based on a prediction error; according to the method, objective and tampering-resistant power data can be fully utilized, the anomaly detection performance of the non-stationary sequence is remarkably improved, and a reliable non-intrusive intelligent monitoring solution is provided for coal mine production safety.

The invention discloses a multi-modal noise reduction evaluation model based on density sensitive clustering, and the model proposes a bidirectional optimization framework of dynamic coupling unsupervised feature extraction and supervised dimension reduction for the dimension reduction demand of high-dimensional data before the high-dimensional data is input into a machine learning model. A topological feedback channel between a high-dimensional space and a low-dimensional manifold is constructed, and a comprehensive scoring system of six types of orthogonalization evaluation indexes is combined, so that automatic selection of a dimension reduction algorithm is realized. The method comprises the following steps: designing a cross-space structure fidelity verification mechanism based on bidirectional topological mapping, a dynamic range calibration and robust normalization system of a multi-modal index, an adaptive search and quality constraint engine of a density clustering parameter, and a composite scoring function design oriented to resource optimization. The method is suitable for a high-noise multi-modal data scene formed by multi-dimensional data including spectral wavelength, intensity, pulse width and the like and different mathematical quantities, all dimensions of the multi-dimensional data can be associated with different databases, and potential association of statistical or physical significance exists among the data.

The invention relates to the technical field of point cloud processing, and discloses a road point cloud semantic segmentation method based on an orthogonal pyramid double-attention network, and the method comprises the steps: introducing an enhanced double-attention orthogonal branch module into an encoder network, and carrying out the decoupling and fusion of local geometric features and global semantic features through dynamic orthogonalization processing; then, through a cross-layer multi-scale information interaction module, features of each level of the encoder are aggregated into a multi-scale fusion feature as global context prior; in a decoder network, a pyramid gating fusion module is utilized to intelligently fuse context features of a decoding path and detail features of an encoder under the guidance of global context priori. According to the method, the problems of redundant feature expression, insufficient cross-layer information interaction and low decoding detail recovery precision in the prior art are solved, the precision and robustness of semantic segmentation of the model in a complex road scene are improved, and the effects of distinguishing similar objects and reconstructing object boundaries are remarkable.

The invention relates to the technical field of pipelines, in particular to a complex FLNG cabin three-dimensional pipeline layout method which comprises the following steps that S1, basic parameters are initialized; s2, calculating initial solutions by adopting an intelligent initial strategy based on obstacle density, and taking an optimal solution in the initial solutions as a current global optimal solution; s3, entering an iteration process, calculating an energy factor, if the energy factor is greater than 1, performing long-distance migration with the probability of 30%, and performing hole digging with the probability of 70%; otherwise, randomly foraging or avoiding the predator at equal probability; s4, periodically performing an elitism and neighborhood search strategy; s5, executing an adaptive orthogonal constraint relaxation strategy and boundary limitation operation; and S6, calculating and determining a final global optimal solution, and outputting a final layout result. According to the method, the calculation efficiency and the convergence speed are improved, the actual engineering situation is fit, and the engineering cost is reduced on the basis of preferentially meeting orthogonalization feasibility.

The invention relates to the technical field of physical confrontation camouflage, in particular to a gradient calibration confrontation texture generation method and device, equipment and a storage medium. Comprising the steps of obtaining a grid model and a plurality of view angle parameters; rendering the grid model based on the current view angle parameter, and generating texture coordinate mapping and a foreground mask; sampling a to-be-optimized texture image based on texture coordinate mapping to obtain a target area image, and synthesizing an input analog image through a foreground mask; calculating a total loss value through the target detection model, and obtaining an initial gradient; optimizing the texture image based on the initial gradient division and executing gradient calibration; performing ascending sorting according to the total loss value of each view angle, performing orthogonalization on all gradients to remove redundant gradient components, and generating a fusion gradient; and updating the to-be-optimized texture image based on the fusion gradient, repeatedly executing until a convergence condition is met, and outputting an adversarial texture image. According to the method, the stability and the effectiveness of the attack texture can be improved under the multi-view and multi-scale conditions.

The invention discloses a bus duct joint deterioration early warning method based on machine learning and causal inference, and aims to solve the problem of indirect estimation under the condition of no direct contact resistance measurement. According to the invention, through the event-driven neural tool variable modeling and orthogonalization constraint, the energy conservation and environmental boundary coding multi-channel thermoelectric physical information neural network, the contact resistance is only coupled to a current square heating item, and the weight-shared anti-fact gating generates baseline temperature rise and residual causal decomposition. And a risk score is constructed by fusing model parameters and extrapolation uncertainty, so that robust estimation and interval early warning of the contact resistance and the degradation rate under a complex environment and load disturbance are realized, and the early warning accuracy and reliability are improved.

The invention relates to the technical field of computer-aided analysis, in particular to a floating foundation structure dynamic response real-time analysis system and method.The system comprises a residual vector calculation module for generating a residual force vector based on a full-order external excitation force vector, an increment base generation module for calculating a modal amplitude and constructing an increment base vector, and an increment base generation module for generating an increment base vector; the reduced-order matrix updating module constructs an enhanced reduced-order model by using the updated projection matrix, and the transient response solving module solves and obtains a corrected generalized coordinate value based on the enhanced reduced-order model and maps the corrected generalized coordinate value into a dynamic displacement field. According to the method, a residual vector is generated by calculating a difference value between full-order external excitation and internal force, residual force is mapped into static deformation distribution by using a stiffness matrix inverse operator, an incremental base is constructed through orthogonalization processing, adaptive updating of a projection space of a reduced-order model is realized, the system dimension is reduced, and the precision is guaranteed. And the calculation efficiency and the real-time performance of dynamic response analysis of the floating structure are improved.

The invention discloses a mathematical modeling AI prediction system in financial risk assessment, which relates to the field of financial data processing and comprises a feature principal component extraction module, a compensation alignment execution module, a momentum regulation and control decision module, a gradient direction optimization module and a risk assessment output module. According to the method, principal components are extracted through covariance matrix decomposition of the qualification-debt ratio and the cash flow fluctuation ratio, feature redundancy is eliminated, the core risk characterization efficiency is improved, a cross-market compensation matrix of transaction frequency and price fluctuation is constructed in combination with SVD and cosine similarity, and the model environment adaptability is enhanced. And analyzing momentum attenuation by adopting a dynamic sliding window of a second derivative matrix eigenvalue ratio, and establishing a regulation and control mechanism to suppress gradient oscillation. Orthogonal interference of historical momentum vectors is eliminated through Schmidt orthogonalization projection, and the accuracy of the gradient optimization direction is guaranteed. And converting abstract numerical values into interpretable risk level classification by using a hyperbolic tangent function in combination with credit rating interval mapping.

The application relates to the technical field of electric digital data processing, and discloses a source code security vulnerability semantic detection method based on a large language model, which comprises the following steps: analyzing to-be-detected source code to extract execution path constraint features, and constructing an orthogonal characteristic basis vector sequence through orthogonal feature extraction; inputting the to-be-detected source code into a large language model to obtain an initial semantic tensor; performing orthogonal projection on the initial semantic tensor to a code security constraint subspace constructed by the orthogonal characteristic basis vector sequence, and performing weighted aggregation in combination with attention weight distribution information entropy to generate a refined semantic vector; and calculating the vector deviation distance between the refined semantic vector and a preset vulnerability feature distribution center; the application uses a logical subspace projection mechanism to eliminate business semantic noise, establishes the association between the detection conclusion and the code execution logic, and improves the precision of identifying hidden logical offset vulnerabilities.

The invention relates to the field of sentiment analysis, in particular to an aspect sentiment triple extraction method based on double-feature enhancement of a graph neural network, which comprises the following steps of: inputting a text X for encoding and splicing, and outputting features through dimension conversion; inputting the features into a graph convolution decorrelation module to perform feature enhancement of a graph structure, and expanding the features of nodes by using an orthogonalization method; a syntactic structure and topological distance information of the graph are fused into a Transform module to serve as a graph neighbor enhancement module, and enhanced term features are obtained; and constructing mark information of text words, and decoding term features by using a span reasoning algorithm to obtain a final aspect emotion triple. The method has the advantages that a double-feature enhancement mechanism is introduced into a graph neural network framework, the problem of balance between insufficient syntactic information utilization and weak semantic modeling ability of a traditional aspect emotion triple extraction method is effectively solved, the over-smoothing problem is relieved, and the information extraction ability is remarkably enhanced.

The invention provides a project task popularity assessment method and device, and the method specifically comprises the following steps: (1) splitting a project task text into a multi-task description text, and extracting task keywords and feature vectors of all tasks; (2) carrying out orthogonalization processing on the feature vectors of the tasks; (3) performing intensity standardization on the feature vector after orthogonalization processing; and (4) predicting the activeness of the corresponding task based on the orthogonalized feature vector and the intensity-standardized feature vector. According to the method, the activeness of the task is predicted according to the dynamic characteristics of the task, the progress change of the task can be responded in real time, and the adaptability and the flexibility in multiple types of projects are improved.

A state estimation method for mobile robot navigation is provided which includes: step S1, initializing the processing module, and then controlling the field-programmable gate array to obtain measurement values of a moment k collected by the inertial measurement unit and the localization apparatus in a constant sampling period; step S2, controlling the application-specific integrated circuit to send the measurement values to the processing module; step S3, controlling the processing module to, based on the measurement values, sequentially perform vector padding, additional inertial vector acquisition, normalization processing, alignment error computation, filter measurement value computation, bias estimation value computation, velocity estimation value computation, pose estimation value computation and orthogonalization processing to obtain state estimation values of the moment k; step S4, controlling the processing module to send the state estimation values to the human-machine interface for visual display.

The invention relates to the technical field of bearing fault detection, in particular to a bearing fault sparse feature extraction method based on local feature online learning, and the method comprises the steps: collecting a bearing vibration original signal of warehouse logistics equipment, and carrying out the normalization processing; performing sliding window segmentation on the processed bearing vibration signal based on a kurtosis index, and selecting a preset number of signal segments with the maximum kurtosis as an initial atom set; carrying out orthogonalization operation on atoms in the initial atom set in sequence to form an online learning dictionary based on local feature learning; carrying out convolution operation on atoms in the online learning dictionary and the normalized bearing vibration signal to obtain a sparse coefficient matrix, and executing soft threshold operation; kurtosis values of sparse coefficient vectors in the sparse coefficient matrix are calculated respectively, the sparse vector with the maximum kurtosis value is selected as a target feature vector for envelope spectrum analysis, and finally whether the bearing of the warehouse logistics equipment breaks down or not is judged through an envelope spectrum.

The invention relates to the field of data processing, in particular to an advertisement click passing rate prediction method and system. The method comprises the following steps: generating a sparse feature vector according to original feature data; inputting the sparse feature vector into a shared embedding layer to obtain a feature embedding matrix; carrying out weighted distribution on the feature embedding matrix by utilizing the explicit branch weight and the implicit branch weight to respectively obtain an explicit input vector and an implicit input vector; extracting to obtain an explicit interaction feature vector; extracting to obtain an initial implicit interaction feature vector; mapping the explicit interaction feature vector into an alignment feature vector, and calculating a projection component of the initial implicit interaction feature vector in the direction of the alignment feature vector; an orthogonalization implicit interaction feature vector is obtained; and generating a prediction result of the advertisement click passing rate. By adopting the method provided by the invention, the accuracy and generalization ability of the advertisement click passing rate prediction result can be effectively improved.

An apparatus for use by a coherent receiver is configured to perform obtaining a first set of coefficients indicating a first filter in an adaptive equalizer of the coherent receiver, wherein the first filter is an adaptive Multi-Input Multi-Output, MIMO, filter with N taps in each branch, wherein N is an integer number; determining a further set of coefficients indicating a MIMO filter with a single tap in each branch based on the first set of coefficients; determining a first matrix based on the further set of coefficients; orthogonalizing the first matrix, thereby obtaining an orthogonalized matrix; determining a state-of-polarization based on at least part of the orthogonalized matrix.

The invention discloses a rapid calculation method and device for an energy maximum tracking model, and relates to the technical field of demand analysis, and the method comprises the following steps: dividing an original training set into a training set and a test set, the training set being used for modeling and parameter adjustment, the test set being used for evaluation, and preprocessing the training set; correlation analysis is carried out on training set features, local correlation measurement is used to avoid multiple collinearity, and an energy value F of each feature is calculated; according to the method, the multi-collinearity problem is avoided by calculating the energy value of each feature and selecting the feature with good prediction, redundant features can be identified and removed through local correlation measurement, the accuracy of feature selection and the robustness of the model are improved, orthogonalization processing helps to eliminate the dependency relationship between the features, and the robustness of the model is improved. The introduction of the regularization item is helpful for controlling the complexity of the model, avoids the overfitting caused by the too complex model, and obtains a stable and accurate prediction result.

The invention provides a blind noise reduction method for MEMS multi-sensor self-contrast learning, which belongs to the technical field of MEMS, and comprises the following steps: receiving observation data of MEMS multi-sensors, and constructing a whitening matrix to obtain whitening data; initializing a separation matrix by using a random unit vector, solving an optimal separation vector, and outputting a normalized signal source and a noise source after orthogonalization processing; enabling a single observation signal, a signal source and a noise source to share encoder weight extraction features, and inputting a decoder to reconstruct a signal; and calculating the total loss, and outputting a high-precision noise reduction signal when the total loss is minimum. According to the blind noise reduction method for MEMS multi-sensor self-contrast learning, the difficulty of non-Gaussian modeling for unknown environment interference in a traditional method is solved, and the problem of amplitude uncertainty of signal blind separation under the state that a physical quantity to be measured does not have prior information is solved.

The invention discloses a multi-user hidden networking method based on double time scale constraints. The method comprises the following steps: 1, slicing time-frequency network resources to obtain an available time-frequency slot set; 2, obtaining a basic random sequence M < k > under the action of a system clock and a user authentication key; 3, performing cyclic shift transformation on the basic random sequence M < k > to obtain an extended sequence family NL * K; step 4, carrying out orthogonalization processing on the extended sequence family NL * K to obtain a decision sequence family HL * K, and guiding multiple users to access the network; 5, authorized user access protocol performance analysis is carried out, and dual-time-scale constraint is carried out to achieve the optimal networking throughput; step 6, reliability and security analysis is carried out on the covert communication system; and 7, analyzing the tolerance of the covert communication system, and solving by using an improved SA-PSO algorithm. According to the method, multiple users are guided to asynchronously access the covert communication network through dynamic management and allocation of spectrum resources, and the method has the characteristics of high efficiency, flexibility, stability and accuracy.

The present invention discloses a method and system for analyzing electromagnetic scattering characteristics of metal targets based on a hybrid nested equivalent source direct solver. Specifically, the method comprises the following steps: establishing a target equivalent surface current model to be solved, and dividing the global basis functions to be solved corresponding to the model into multiple groups in different spatial regions; constructing an elimination matrix for the far-field coupling matrix blocks for each group to compress and eliminate the admissible blocks; eliminating the inadmissible blocks by calculating the Schur complement of the remaining inadmissible matrix blocks; compressing the fill-in matrix blocks added to the admissible blocks, extracting their column vectors and adding them to the orthogonalized receiving matrix of the next group; skipping the remaining decomposition process of the group in the event of overflow caused by matrix addition or translation from a sublayer to a parent layer; and sequentially traversing each layer and group to solve the system matrix equation, ultimately obtaining the electromagnetic scattering characteristics of the target, such as the radar cross-section. The method has high computational efficiency and lower memory requirements.

The invention relates to an image confrontation sample generation method based on key region constraint, which belongs to the technical field of computer vision, and comprises the following steps of: screening a corresponding target image through a target model classification result, identifying by a proxy model to obtain a tail end feature graph and a full connection layer weight, and calculating a category activation mapping graph; the method comprises the following steps: segmenting a target image into a plurality of target areas through instance segmentation, screening candidate segmentation areas from each target area by using a category activation mapping graph, identifying a covered image obtained based on the combination of the candidate segmentation areas by using an agent model, and screening out key areas; the method comprises the following steps of: sequentially performing gradient-free optimization based on Schmidt orthogonalization on each key region, updating a disturbance vector according to an output result of a target model, and finally generating an adversarial sample after multiple rounds of iteration, thereby solving the problems that the efficiency of generating an image adversarial sample in the prior art is low, the process is complicated, and the concealment and the disturbance effect cannot be considered at the same time.

The present application relates to a multi-view clustering method, apparatus, device and storage medium based on latent variables. The method comprises: obtaining latent variables of multi-view data, obtaining latent representations of samples according to the latent variables; using V A neural network will N The latent representation of each sample is mapped to the feature space of the view data to obtain the reconstructed data representation of each view data. The first loss is constructed based on the distance between each reconstructed data representation and the corresponding view data. The values ​​of the latent variables are determined based on the latent representation. The values ​​of each latent variable are orthogonalized and constrained to have a modulus of 1 to obtain the second loss. The linear kernel matrices of each view data and the corresponding reconstructed data representation are kernel-aligned, and the alignment losses of each view data are weightedly fused to obtain the third loss. A global loss function is derived from these losses. The global loss function is optimized, and the optimized latent representation is used for multi-view clustering. This method can effectively improve the clustering performance of multimodal data.

This application relates to the field of point cloud processing technology and discloses a road point cloud semantic segmentation method based on an orthogonal pyramid dual-attention network. This method introduces an enhanced dual-attention orthogonal branch module into the encoder network, utilizing dynamic orthogonalization to decouple and fuse local geometric features and global semantic features. Then, through a cross-layer multi-scale information interaction module, features from each layer of the encoder are aggregated into multi-scale fused features as a global context prior. In the decoder network, a pyramid-gated fusion module is used to intelligently fuse the contextual features of the decoding path with the detailed features of the encoder under the guidance of the global context prior. This invention solves the problems of redundant feature representation, insufficient cross-layer information interaction, and low accuracy in decoding detail recovery in existing technologies, improving the accuracy and robustness of semantic segmentation in complex road scenes, and showing significant effects in distinguishing similar objects and reconstructing object boundaries.

The invention provides a self-interference elimination model generation method and device, network equipment and a storage medium, and is applied to the technical field of wireless. The method comprises the following steps: establishing a self-interference elimination model according to a signal to be processed, and obtaining at least one primary function of the self-interference elimination model; carrying out orthogonalization processing on the at least one primary function through the GPU to obtain a primary function after orthogonalization processing; storing vectors of the basis function after orthogonal processing in the GPU, and constructing a first vector table; and transmitting the first vector table to a CPU, and performing parameter learning on the first vector table through the CPU to obtain model parameters of the self-interference elimination model. By adopting the method, the orthogonalization processing and the parameter learning process in the self-interference elimination are respectively processed in the GPU and the CPU, so that the self-interference elimination is carried out by utilizing hardware resources to the maximum extent, the processing efficiency of the self-interference elimination is improved, and the method can adapt to computing power demand scenes with large bandwidth and high data rate.