17 results about "Normalized mean square error" patented technology

The invention discloses a photon terahertz communication system channel modeling method based on a conditional generative adversarial network, and belongs to the field of terahertz communication. The method specifically comprises the following steps: firstly, building a photon terahertz communication system comprising a signal transmitting module and a signal receiving module; after symbol sequences of a transmitting end and a receiving end are preprocessed, a training set and a test set are divided. Then, a conditional generative adversarial network used for channel modeling of the photon terahertz communication system is constructed, the conditional generative adversarial network comprises a generator and a discriminator, and a self-attention mechanism is introduced into the middle layer of a multi-layer full-connection network of the generator and discriminator network; training the conditional generative adversarial network by using the training set to obtain optimized parameter configuration; and finally, inputting the condition vector and the random vector of the test set into a generator to obtain generated data, and performing evaluation by using a normalized mean square error. According to the method, a self-attention mechanism is introduced in a scene with a severe channel condition, so that the robustness and the stability of modeling can be remarkably improved.

The invention discloses an electric power traffic coupling network sentry node identification method and system, and the method comprises the steps: carrying out the value quantification and screening of collected electric power traffic multi-source heterogeneous data based on an optimized gradient value, and outputting high-value data; the method comprises the following steps: constructing a unified hypergraph model of an electric power traffic coupling network, defining electric power nodes, traffic nodes and charging station nodes as a heterogeneous node set, constructing multiple types of hyperedges according to physical connection, flow similarity and a cross-network supply-demand relationship, and extracting a node high-order topological feature embedding matrix by using a hypergraph convolutional neural network; based on a reinforcement learning algorithm, a sentinel node identification problem is modeled as a Markov decision process, and a sentinel node set enabling a normalized mean square error function to be minimum is dynamically searched and output. According to the method, in a disaster environment with high data noise and limited computing resources, the local key nodes are utilized to accurately invert the guard node deployment of the whole-network macroscopic state.

The invention provides a channel estimation method, device and equipment, a storage medium and a program product, which are applied to the technical field of wireless communication. The method comprises the following steps: collecting state information of a current communication environment; inputting the state information to the pilot frequency scheduling strategy network to obtain a pilot frequency activation mask, wherein the pilot frequency activation mask is used for indicating whether the user terminal sends the pilot frequency or not; constructing a sparse pilot frequency matrix based on the pilot frequency activation mask, and transmitting pilot frequency signals according to the sparse pilot frequency matrix to generate a pilot frequency observation matrix; the pilot frequency observation matrix and the pilot frequency activation mask serve as input of a channel estimator, missing channel information is complemented through back diffusion iteration, a channel estimation result is obtained, and the channel estimator is constructed based on a generation diffusion model; and calculating a normalized mean square error of the channel estimation result and a real channel, constructing a reward function in combination with the normalized mean square error and the pilot frequency usage amount, and updating a scheduling strategy of the pilot frequency scheduling strategy network based on the reward function.

The invention relates to a large-scale MIMO (Multiple Input Multiple Output) key generation method based on channel mapping and compression quantization, which specifically comprises the following steps of: aiming at the problem that the channel mapping accuracy of the existing model is insufficient, providing a CoSTMNet model based on a hybrid codec, which is used for modeling a channel mapping function which cannot be mathematically solved and improving the channel mapping accuracy; in order to solve the problem of insufficient key consistency of the existing FDD key scheme, the invention provides a double-bit quantization algorithm PCGQ based on partition compression and gray code coding, and designs a full-process physical layer key generation scheme in an FDD mode in combination with a CoSTMNet channel mapping model. Compared with a comparison object, the algorithm scheme provided by the invention has better performance in the aspects of normalized mean square error, Pearson correlation coefficient, key consistency and other indexes.

The invention particularly relates to a heuristic global optimization algorithm-based multi-dimensional nonlinear behavior model establishment method, which comprises the following steps of: changing multi-dimensional nonlinear characteristic equipment of a target component, and acquiring multi-dimensional nonlinear characteristic data of the target component; selecting a behavior model structure type; setting a target function and a constraint function based on the behavior model structure type, and searching an optimal behavior model structure with the minimum fitness value by using a heuristic global optimization algorithm; the normalized mean square error of the optimal behavior model structure is compared with a set threshold value, and whether the calculation result meets the threshold value requirement or not is judged; and if a threshold requirement is met, outputting the current behavior model structure and the parameters as a final behavior model, otherwise, repeating the step of selecting the behavior model structure type to the step of outputting the behavior model, and re-establishing the behavior model. According to the method, the accurate multi-dimensional nonlinear behavior model with the globally minimum mean square error is realized.

The invention discloses a millimeter wave large-scale MIMO adaptive-fast convergence sparse recovery algorithm, which is an adaptive multi-dimensional joint correlation synchronous orthogonal matching pursuit algorithm. According to the algorithm, a plurality of atoms are selected in parallel in single iteration, so that the convergence process is remarkably accelerated. Moreover, an adaptive termination criterion can be realized without prior information such as channel sparseness and noise power. Simulation results show that the convergence speed of the algorithm is increased by about 50% compared with that of a traditional SOMP algorithm, and a normalized mean square error equivalent to that of the traditional SOMP algorithm can be obtained under the condition of lacking channel sparseness and noise power prior information. Especially in an environment with a signal-to-noise ratio of 8-23dB, even if channel noise power information does not exist, sparseness estimation errors can be effectively reduced, and the method is superior to a traditional algorithm. Therefore, the algorithm provided by the invention has strong self-adaptability and can effectively cope with the dynamic change of the channel.

The application provides an R2* image synthesis method based on a generative adversarial network, comprising the following steps: S1, data preprocessing: calculating and correcting R2* values from MEGRE sequences; S2, ROI segmentation: after matching the R2* graph with the AALv3 template, the average R2* values of the regions of interest are extracted from the synthesis graph and the real graph; S3, GAN model: the generator inputs the T1 weighted image and the T2 weighted image, and generates the corresponding R2* image; the discriminator distinguishes the synthesized R2* image generated by the generator from the real R2* image obtained from the real MEGRE sequence; S4, quantitative evaluation of the generated image: the normalized mean square error, the peak signal-to-noise ratio and the structural similarity index are used to evaluate the similarity between the synthesized image and the real R2* graph; S5, statistical analysis: the R2* values of the regions of interest in the synthesized graph and the real R2* graph are analyzed and compared. The application realizes the auxiliary diagnosis of neurodegenerative diseases similar to Parkinson's disease from the perspective of medical image synthesis and processing.

The invention discloses a multi-IRS channel prediction method based on a meta-learning algorithm and related equipment, and relates to the field of intelligent reflector auxiliary communication, and the method specifically comprises the steps: obtaining an uplink cascade channel, and forming a real and virtual dual-channel tensor; establishing a two-dimensional convolutional network to realize uplink-to-downlink mapping; a phase consistency item is constructed based on phase unwrapping, and the phase consistency item, a mean square error and weight regularization form composite loss; executing meta-training by taking a plurality of subsystems as source tasks, and performing joint optimization in support set inner layer updating and query set outer layer updating; and carrying out rapid self-adaption on the target subsystem by using a small number of labeled samples, outputting a downlink cascade channel and evaluating a normalized mean square error. According to the scheme, stable prediction is realized under the condition of few samples, and engineering deployment is facilitated.

The invention relates to a large-scale MIMO (Multiple Input Multiple Output) key generation method based on joint optimization and metric learning, which specifically comprises the following steps: an RIS (Remote Information System) and BS (Base Station) joint alternative optimization algorithm is used for optimizing the signal reachable rate of the whole system and providing a reliable channel environment for key generation between legal communication parties; a multi-scale CNN and ViT hybrid network model based on a metric learning twin network architecture is used for extracting deep embedded reciprocity features in high-dimensional channel information of a legal communication party in a large-scale MIMO system, extraction of local features and modeling of global dependence can be considered at the same time, and a high-consistency data basis can be provided for key generation. A full-process physical layer key generation scheme in a TDD mode is designed in combination with a proposed RBAO joint optimization algorithm and a hybrid network model, and the method has excellent performance in indexes such as normalized mean square error, Pearson correlation coefficient, key consistency and the like.

The present application belongs to the technical field of power quality measurement, and provides a harmonic distortion type identification method and system based on symmetry characteristics, which obtains power signal data according to a preset sampling window sliding; processes the obtained data, extracts harmonic components, calculates single-window harmonic energy variance indexes, performs grouping processing, compares with a preset center-symmetry normalized mean square error threshold, judges whether each group of data conforms to center-symmetry distribution, and then compares with an error threshold of inter-harmonic characteristics to obtain an inter-harmonic criterion variable; processes, calculates and analyzes time domain data of any complete single period, compares the consistency degree of the translated first half cycle data and the inverted second half cycle data to obtain an even harmonic criterion variable; and comprehensively identifies the signal harmonic distortion type according to the inter-harmonic criterion variable and the even harmonic criterion variable. The present application realizes accurate identification of the power signal harmonic distortion type.

This invention discloses a downlink channel estimation method based on a smart reflector-assisted MIMO communication system. It constructs a downlink system model of a smart reflector-assisted large-scale MIMO communication system; constructs a row sparse matrix and substitutes it into the downlink system model, thereby transforming the smart reflector-assisted large-scale MIMO downlink concatenated channel estimation into a standard row sparse recovery problem; it uses an off-network sparse Bayesian algorithm to iteratively update the noise accuracy, the accuracy vector of E, and the off-network gap parameter; it sets an iteration update termination condition; it sets a threshold and uses this threshold to select the effective angle set of the channel; it obtains accurate channel state information based on the effective angle set; and it uses the normalized mean square error (NMSE) to determine the accuracy of the calculated concatenated channel. This invention solves the problem of reduced output signal-to-noise ratio causing performance loss in channel estimation caused by the decoupling process of converting concatenated channel estimation into a sparse signal recovery problem through pseudo-inverse operations in existing technologies.

The invention belongs to the technical field of ultrasonic imaging, and discloses a matrix imaging method based on a Laplacian operator, which comprises the following steps: performing feature extraction on an original image by using the Laplacian operator to obtain feature information; obtaining a dynamic normalized mean square error between the original image and the reconstructed image obtained based on the feature information; judging whether a dynamic normalized mean square error calculation result is within a preset range or not; and if yes, performing full-focusing imaging processing on the corresponding original image to obtain a target image. According to the method, the intermediate image data is evaluated once before imaging, the original image is selectively imaged, and the overall calculation amount of ultrasonic imaging is reduced by screening the original image, so that the working efficiency of ultrasonic imaging during living tissue detection is improved.

The present application belongs to the technical field of photonic integrated interferometric imaging, and particularly relates to a kind of optimal spectral channel design method of photonic integrated interferometric imaging system. It includes: S1: according to the spectral distribution of the target to be measured, the interference array of the photonic integrated interferometric imaging system is designed;S2: set the spectral channel number and initial spectral sampling interval of the photonic integrated interferometric imaging system;S3: the spectral distribution of the target to be measured sampled is inversely Fourier transformed one by one, and the reconstructed image corresponding to each spectral sampling interval is obtained;S4: all reconstructed images are normalized;S5: the value of normalized mean square error and the value of normalized peak signal-to-noise ratio are used to determine the optimal spectral channel of the photonic integrated interferometric imaging system. The present application effectively improves the design efficiency of the photonic integrated interferometric imaging system.

The invention relates to an intelligent reflecting surface (IRS) assisted multi-user communication sensing integrated (ISAC) system channel estimation method based on radar sensing priori and structured sparsity, and belongs to the technical field of wireless communication. According to the method, the IRS is deployed to construct a cascade communication link between the base station and multiple users, and the radar sensor is arranged near the IRS to realize perception blind compensation, so that the problems of communication interruption and radar blind areas of a traditional ISAC system in a shielding environment are solved. The core technology content comprises the following steps: revealing the partially overlapped structured sparsity existing in the IRS-assisted multi-user communication process, and providing theoretical support for low-overhead channel estimation; a typical user + remaining user two-stage channel estimation framework is provided; an angle information conversion model between the sensor and the IRS is constructed, the problem that angle coordinates of a radar target relative to the sensor and the IRS are inconsistent is solved, and perception-communication cross-domain information fusion precision is improved. According to the method, the model design is complete, the algorithm is reasonable, the normalized mean square error and pilot frequency overhead of channel estimation can be remarkably reduced, meanwhile, the radar angle estimation precision is improved, and the method is suitable for a B5G / 6G multi-user ISAC complex shielding scene.

The application discloses a channel map construction method based on a graph neural network, relates to the technical field of wireless communication and artificial intelligence, and comprises the following steps: firstly, based on observed channel data, an initial channel matrix is generated for unobserved nodes through pre-interpolation; secondly, a spatial graph structure is constructed by using a Wasserstein distance; then, an inductive aggregation graph neural network is adopted to realize spatial completion and smooth reconstruction of channel data through hierarchical neighbor sampling and feature aggregation; finally, a combined loss function composed of a reconstruction loss and a smooth loss is introduced to optimize model parameters, and an accurate and continuous channel map is obtained. Compared with traditional interpolation or deep learning methods, the application can significantly improve channel completion accuracy and spatial consistency, and experimental results show that the root mean square error of the completed channel and the measured channel is reduced to 2.72, and the normalized mean square error is only 0.0054, so that the application has the advantages of high precision and high robustness and can be widely applied to 6G network planning, channel prediction and intelligent resource management and the like.

The present application relates to the technical field of electromagnetic calculation, in particular to a kind of electromagnetic inverse scattering imaging method based on multi-scale coordinate attention mechanism neural network. Including step 1: scattering field data is acquired using transmitting and receiving antenna and initial rough image is generated by back propagation algorithm;Step 2: neural network model is constructed, which is composed of input layer, multi-scale feature extraction module, adaptive fusion module, coordinate attention module and output layer;Step 3: the cost function of the combination of normalized mean square error and regularization constraint is used to optimize network parameters;Step 4: scattering body reconstruction is carried out by training multi-scale coordinate attention mechanism neural network;The present application makes full use of the prior information of scattering field data and rough image, improves the reconstruction accuracy of high-contrast scattering body edge and details through multi-scale feature fusion and coordinate attention mechanism, and has the advantages of good imaging effect, strong generalization ability and high precision.

The invention discloses an active user detection and channel estimation method based on a score generation model, belongs to the technical field of wireless communication, and is suitable for a large-scale machine type communication scene. The method specifically comprises the steps that a score generation model is firstly constructed, channel prior distribution is learned through channel data offline training, and the model is optimized through an annealing denoising score matching loss function; in the on-line stage, pilot frequency receiving signals and a pre-training model are combined, joint active user detection and channel estimation are achieved through an annealing Langevin dynamics algorithm, and estimation precision is improved through dual-cycle progressive optimization; and determining an active user state based on an estimation result and calculating a channel normalized mean square error.