76 results about "Source localization" patented technology

The invention discloses an anomaly detection and intelligent diagnosis method, system and device based on digital power grid multi-source data and a medium, and belongs to the technical field of anomaly detection, and the method comprises the steps: obtaining multi-source operation data from a power grid operation process, carrying out the preprocessing, and generating a standardized data set; time sequence features are extracted based on historical data, a power grid state reference model is constructed, and normal operation states in different load scenes are represented; on the basis of deviation calculation of the standardized data set and the power grid state reference model, abnormal candidate signals are detected, and high-confidence-coefficient abnormal signals are screened and generated; determining an abnormal source based on the high-confidence abnormal signal in combination with a power grid topological structure, and performing analysis to obtain fault type information; and generating a control instruction according to the fault type information and issuing the control instruction to a power grid control system. According to the method, a complete technical scheme of multi-dimensional data fusion, dynamic deviation detection, high-confidence anomaly screening, anomaly source accurate positioning and fault type rapid diagnosis is realized.

The invention discloses a multi-modal deep learning traceability method and system fusing magnetoencephalogram and electroencephalography, and relates to the technical field of artificial intelligence and neuroimage.Real magnetoencephalogram signals and electroencephalography signals are preprocessed and then input into a traceability model, the probability of occurrence of a source in a corresponding area is predicted, and the traceability of the source in the corresponding area is obtained by combining an imported source partition distance matrix. A final traceability result is obtained; the training process of the traceability model is as follows: constructing a generative adversarial network, and generating a multi-modal neural electrophysiological data set; inputting the multi-modal neural electrophysiological data set into a residual network of a double-branch structure, and performing stage hierarchical extraction and decoupling on magnetoencephalogram signals and electroencephalogram signals respectively; extracting features in different stages by using a multi-scale convolution module, fusing the extracted features, inputting the fused features into a classifier, defining a loss function, and updating trainable parameters of the traceability model; the traceability method improves the accuracy and generalization ability of traceability positioning.

An autonomous system for detecting, localizing, and potentially deactivating chemical threats or emissions using multiple sensing modalities and reinforcement learning techniques. The system includes visual sensors (e.g., RGB, RGBD, LIDAR), non-visual sensors (e.g., gas concentration, airflow, GPS, RADAR), a neural network architecture and processor to fuse information from different sensors, a module based on deep reinforcement learning for decision making, and a robotic interface for executing actions. The neural network extracts relevant information from sensor streams and encodes them into a joint embedding space. The module considers the current observations, historical data, and previous actions to determine the optimal action for threat localization under partially observable conditions. The system is trained in simulated environments to minimize source localization time while accounting for various constraints. The autonomous system enables effective chemical threat detection and source localization in complex, dynamic environments without endangering human operators.

The invention relates to the field of hydraulic structures, in particular to a three-dimensional damage rapid positioning method for a complex aqueduct structure. Acoustic emission or microseismic signals in the aqueduct structure are collected, and the first arrival time of the signals is picked up through an STA / LTA long and short time window energy ratio method; constructing a curved surface structure model composed of a groove surface and a side wall; parameterizing the cross section, and constructing a groove surface propagation path set according to the position relationship between the damage point and the sensor in a parameter space; for each pair of damage points and sensors, selecting a path with the minimum groove surface path length as a shortest propagation path, and calculating theoretical arrival time according to the shortest propagation path; establishing a residual function between the theoretical arrival time and the actually measured arrival time, and constructing a jacobian matrix of arrival time change; and solving the coordinates and occurrence time of the damage source. According to the method, the parameterized model of the aqueduct section is established, traditional ray tracing is replaced with the curved surface shortest path thought, and the numerical difference and rapid iteration methods are combined, so that efficient and robust damage source positioning is achieved.

The invention discloses a wide-area oscillation source positioning method and system based on physical feature embedding and a graph attention network, and the method comprises the steps: obtaining synchronous phasor data of a power system, carrying out the modal parameter identification of subsynchronous oscillation and / or super-synchronous oscillation based on the synchronous phasor data, constructing a topological graph of the power system, and carrying out the positioning of a wide-area oscillation source. And constructing the oscillation amplitude, the attenuation factor and the phase of each node obtained by identification into a feature vector of each node in a topological graph, inputting the topological graph and the feature vector into a graph attention network, and outputting an oscillation source positioning result, so that high-frequency recording data after an accident does not need to be waited, and the positioning accuracy of the oscillation source is improved only on the basis of synchronous phasor data. The identified modal parameters are used as physical characteristics to be embedded into the graph attention network, and the interpretability of a physical mechanism and the graph structure mining capability of deep learning are fused, so that subsynchronous oscillation and super-synchronous oscillation are accurately positioned.

The present disclosure provides an interference source localization method and apparatus, a storage medium, and a program product. The interference source localization method comprises: sending a sensing reception configuration to a first network node; receiving interference measurement information fed back by the first network node and determined on the basis of the sensing reception configuration; and determining, on the basis of the interference measurement information, whether an interference source is present, and localizing the interference source.

The invention provides a multi-sensor fusion positioning method and device of a vehicle, equipment and a storage medium, and relates to the field of data processing. The method comprises the following steps: acquiring different and independent multi-source positioning information through a plurality of sensor sub-modules; sorting the multi-source positioning information according to the size of the observation timestamps, and screening the multi-source positioning information based on a preset time delay window; performing filtering fusion calculation on the reserved multi-source positioning information to obtain first positioning state data after time delay; performing time adjustment on the time-delayed first positioning state data by using the motion observation information to obtain second positioning state data at the current system moment; and post-processing the second positioning state data to obtain stable and smooth third positioning state data.

This invention discloses a GNSS interference source localization method based on ADS-B data. The method involves acquiring ADS-B data, saving data points affected by GNSS interference, and, if no NIC is available, constructing ADS-B data with a NIC. The message data is then sorted by ICAO and timestamp. A diagonal matrix W is calculated based on the ICAO, timestamp, NIC, a table showing the correspondence between NIC and received interference power, and the hyperparameter value τ. A residual vector R is calculated based on the message data and the optimization vector, and the optimized vector is then output. This invention analyzes large-scale aviation ADS-B data, sorts and classifies the original message data to determine whether flights are subject to GNSS interference, and abstracts the interference source localization problem into a least-squares problem. An optimization function is constructed, and an iterative optimization algorithm is used to calculate the possible GNSS interference source locations and transmission power, thus more efficiently locating GNSS interference sources.

The invention relates to the technical field of electroencephalogram signal monitoring and processing, in particular to a user category determination method based on an auditory brainstem reaction normal form and source localization, and the method comprises the following steps: obtaining auditory brainstem reaction data and electroencephalogram data when a target user is stimulated by external sound; based on a feature extraction module, feature extraction is carried out on source estimation data corresponding to the auditory brainstem response data and the electroencephalogram data; performing feature fusion on the basis of the features extracted from the auditory brainstem reaction data and the features extracted from the source estimation data, and determining fusion features; and determining the category of the target user based on the fusion features and a user category prediction model. According to the scheme for carrying out early diagnosis and screening on the autism infants based on the auditory brainstem reaction data and the electroencephalogram data, early autism (ASD) screening of the infants can be finally achieved under the support of the electroencephalogram equipment and auditory brainstem reaction testing.

This invention discloses a broadband oscillation disturbance source localization method based on compressed sensing and graph convolutional neural networks (GCNs), comprising two stages: offline training and online localization. In the offline training stage, electrical quantities under subsynchronous / supersynchronous broadband oscillation modes are obtained using measurement data from actual systems or simulation examples, constructing a broadband oscillation offline sample library. Compressed sensing is used to encode and compress the electrical quantities in the offline sample library, obtaining a training set and a test set. The constructed GCN localization model is trained using the training set until the test set reaches the target localization accuracy, resulting in a broadband oscillation localization model. In the online localization stage, a substation collects electrical quantity data and compresses and encodes it using compressed sensing technology; the substation data is uploaded to the master station; at the master station, the feature matrix and system adjacency matrix are input into the trained GCN localization model, and the oscillation source location is output. This invention's method is applicable to broadband oscillation disturbance source localization under various operating conditions.

This invention discloses a non-cooperative signal localization method based on radiation source decoupling, belonging to the field of wireless signal processing and spatial positioning. The method includes: transforming multi-source joint localization into multiple independent single-source localization problems by constructing a source-level decoupling theoretical framework. First, the method utilizes multiple receiving nodes to acquire mixed signals; then, it uses a mask-based separation network (MS-Net) to perform blind source separation on the mixed signals, obtaining the decoupled components of each radiation source. This network does not require prior information such as pilot signals; finally, for each decoupled component, a broadband Doppler model considering time-scale effects is used for direct single-source localization. This invention significantly reduces computational complexity, improves model adaptability and localization accuracy in high-speed motion scenarios, and has good engineering application prospects.

The application discloses a multi-modal deep learning source tracing method and system combining magnetoencephalogram and electroencephalogram, relates to the field of artificial intelligence and neuroimaging technology, and inputs real magnetoencephalogram signals and electroencephalogram signals into a source tracing model after preprocessing, predicts the probability of the occurrence of corresponding regional sources, combines an imported source partition distance matrix, and obtains a final source tracing result; the training process of the source tracing model is as follows: a generative adversarial network is constructed to generate a multi-modal neuroelectrophysiological data set; the multi-modal neuroelectrophysiological data set is input into a double-branch structure residual network to perform stage hierarchical extraction and decoupling on the magnetoencephalogram signals and the electroencephalogram signals respectively; a multi-scale convolution module is used to extract features at different stages, the extracted features are input into a classifier after being fused, and a loss function is defined to update trainable parameters of the source tracing model; and the source tracing method improves the accuracy and generalization ability of source positioning.

The application discloses a kind of deep learning brain electromagnetic source positioning method and system based on residual iteration, comprising: S1, current electromagnetic signal is obtained;S2, based on current electromagnetic signal, obtain preliminary source positioning result;S3, based on preliminary source positioning result, obtain sparse source positioning result;S4, based on sparse source positioning result, obtain maximum intensity activation source;S5, update current electromagnetic signal, repeat S2-S5 until reach preset condition, obtain final source positioning result. Carry out preliminary source positioning using sLORETA, provide a more rough brain source positioning result. Then the result of sLORETA is repaired using neural network, to restore the sparse source of accurate position, and as far as possible restore the intensity of activation source. The application comprehensively sLORETA's fast calculation and preliminary positioning ability, and the high-precision repair capability of neural network, so as to realize more accurate, more stable result in source positioning.

The invention provides a multi-detector radioactive source positioning method and system based on a particle filter algorithm, and the method comprises the following steps: defining a to-be-monitored physical space as a search region, and building a radiation measurement model; randomly generating a particle set in the search area, wherein each particle comprises assumed radioactive source position coordinates and estimated activity parameters; a mobile platform carrying a plurality of radiation detectors is controlled to move in the area, the total count value of actual measurement of each detector is obtained, and the weight of each particle in the particle set is calculated by using the radiation measurement model and the actual measurement values; acquiring an environment background count value, calculating a ratio of an actual measurement value to a background value, and only when the ratio exceeds a preset signal-to-noise ratio threshold value, performing resampling on the particle set; judging whether the particle set meets a positioning ending condition or not, and if yes, determining a radioactive source positioning result by using the particle parameter with the highest weight and outputting the radioactive source positioning result; and if not, taking the weight gathering center as the next target position, and controlling the mobile platform to move towards the next target position.

The invention relates to a radiation source direct positioning method based on random sampling, belongs to the technical field of radiation source direct positioning, and solves the problem of large calculation amount of radiation source direct positioning in the prior art. Comprising the following steps: receiving radiation source signals from different directions by using dispersedly arranged receiving nodes, and sampling received data to obtain observation sample data at different moments; grid division is carried out on the radiation source sensing area coordinate system; randomly selecting a preset number of grid points according to uniform distribution, and calculating likelihood ratio test statistics subjected to time delay compensation as matrix elements of the positions of the grid points based on observation sample data and the positions of the selected grid points; and completing matrix element completion of the whole radiation source sensing area based on a low-rank matrix completion method to form a sensing area receiving signal synchronization likelihood ratio test statistic matrix, and determining a peak position in the sensing area receiving signal synchronization likelihood ratio test statistic matrix as a radiation source position. The radiation source direct positioning method based on random sampling is realized.

A multi-source localization method and system for quadruped robots based on geometric spatial decoupling and full-dimensional dynamic confidence is disclosed. The method involves: acquiring data and inputting it into a data preprocessing module to obtain distortion-free effective point cloud data; inputting the effective point cloud data into a geometric information matrix construction module to construct a geometric information matrix; based on the geometric information matrix, a full-dimensional spatial degradation decoupling and perception module performs subspace decoupling, algebraic stiffness calculation, and temporal inertial smoothing to form a geometric confidence factor in the [0,1] interval; inputting the geometric confidence factor into a multi-source covariance elastic reconstruction module to dynamically reconstruct the observation noise covariance of the lidar and leg constraints according to the geometric confidence factor, obtaining the reconstructed lidar observation constraints and leg contact observation constraints; inputting the reconstructed observation constraints into an error state Kalman filter to output the high-precision pose of the quadruped robot. This invention enables precise localization of quadruped robots.

The application discloses a kind of line array rotary time array shape correction and noise source positioning joint estimation method, system and computing device, and the joint estimation method includes the following steps: preliminary estimation line array element position;Preliminary positioning sound source position;Calculate the delay time of each array element in line array relative to the 1st array element, and calculate the steering vector and array flow pattern matrix;Establish target cost function and constraint condition to optimize and solve the correction amount of each array element in line array;According to the array element position under preliminary estimation array shape and the correction amount of each array element in line array, calculate the corrected line array element position;Based on second-order polynomial fitting method, curve fitting is carried out to each array element position, and the line array element position after curve fitting is obtained;Based on the new array element position, the array flow pattern matrix is corrected, and the sound source position is positioned again under the new array flow pattern using minimum variance distortionless response algorithm.

This invention discloses an underwater electric field source localization method based on a determinant orthogonal propagation operator and adaptive hybrid-driven differential evolution, belonging to the field of electromagnetic positioning and detection. This invention acquires and receives array signals and calculates the covariance matrix. Based on the propagation operator, it constructs a projection operator matrix orthogonal to the array manifold to replace the noise subspace, thereby constructing a determinant spatial spectrum function and defining a cost function. Then, based on this cost function, it constructs an adaptive hybrid-driven differential evolution framework, fusing the alpha evolution algorithm and an adaptive differential evolution algorithm based on successful historical records to solve for the target position coordinates. This invention eliminates the need for eigenvalue decomposition, reducing computational complexity, and accelerates convergence through an adaptive hybrid-driven search strategy, enabling fast and high-precision electric field source localization in complex underwater environments.

This invention discloses a fine motor imagery decoding system based on PINN and multimodal fusion. The system employs a fine motor imagery decoding method based on PINN and multimodal fusion, and includes a multimodal physiological signal synchronous acquisition module, a data preprocessing and correlation analysis module, a dual-stream decoding network module, a physical manifold constraint layer module, an adaptive gradient balance optimization module, and a classification decision module. By deeply embedding biophysical mechanisms into a deep learning framework, this invention addresses the ill-posedness of the source localization inverse problem, overcomes overfitting in small-sample scenarios, and significantly improves the decoding accuracy, robustness, and physiological interpretability of fine hand movements.

This invention discloses a rapid pollution source location system and method based on mobile monitoring data and a source tracing algorithm. The system includes mobile water quality monitoring data collection, data preprocessing, source tracing algorithm model construction, model optimization, pollution source location, and result display. During water quality source tracing using mobile monitoring equipment, a dynamic tracking mechanism is employed. A source tracing algorithm is constructed to enable the monitoring equipment to move autonomously, adapting to changes in the pollution source's location and enabling rapid source location. Considering the energy consumption-accuracy trade-off during the mobile monitoring equipment's navigation, an energy consumption penalty is introduced into the reward function to avoid ineffective movement and improve the source tracing efficiency. Combining historical monitoring data sequences reflects the trend of water quality changes over time, improving the accuracy of model training. This allows for accurate and rapid water quality source tracing, achieving both high efficiency and accuracy.

This invention proposes a radio map construction and non-cooperative radiation source localization method based on agent interaction. The method first utilizes a drone swarm to dynamically sample sparse signals, then uses Gaussian process regression to fuse the signal propagation physical priors to generate a spatial attention weight map. These physical priors are injected as biases into a multi-head attention mechanism, enabling multiple agents to form a consistent joint representation of the radiation source target in a unified semantic space. Next, key semantic features are dynamically compressed and selected through task-driven semantic filtering and interaction. Semantic recovery is performed using masked multi-head self-attention and transposed convolutional networks to suppress noise and output preliminary results for both tasks. Finally, an embodied intelligent feedback closed-loop system is constructed. This system can perceive the environment and task status in real time, dynamically adjust the weights and interaction strategies of both tasks, and store the optimal strategy in long-term memory. This invention improves the accuracy of radio map construction, radiation source localization accuracy, and multi-agent collaborative efficiency in sparse sampling, multi-source aliasing, and dynamic complex electromagnetic environments.

A method for acoustic noise source detection. The method may include disposing an acoustic logging tool into a wellbore, taking a first measurement at a first depth with the acoustic logging tool as the acoustic logging tool traverses down the wellbore, taking a second measurement at a second depth with the acoustic logging tool as the acoustic logging tool traverses down the wellbore, and forming a first noise source localization map based at least in part on the first measurement. The method may further include forming a second noise source localization map based at least in part on the second measurement and combining the first noise source localization map and the second noise source localization map to form a final enhanced noise source localization map.

The invention relates to an electroencephalogram traceability model training method and device, equipment and a medium. The method comprises the following steps: constructing a standardized graph structure data set containing multiple individual electroencephalogram signals, a structure connection group and a source activity true value, and training a graph neural network by adopting a meta-learning framework to extract a common rule of a cross-individual brain connection group and a traceability mapping relationship, so as to form a pre-training model with strong generalization ability; for a new individual, only key parameters associated with connection group features in the model are adjusted through a parameter efficient fine tuning technology, and rapid migration of pre-training meta-knowledge to individual specific connection is realized; finally, while individualized traceability precision is kept, computing resources and data volume required for model adaptation are greatly reduced, the problem of traceability deviation caused by individual brain structure difference in a traditional method is effectively solved, and feasibility and efficiency of an electroencephalogram traceability technology in clinical practice are remarkably improved.

This invention relates to the field of UAV swarm cooperative localization technology, specifically to a UAV swarm cooperative localization method for non-cooperative radiation sources. The method includes: constructing a UAV swarm cooperative localization system model; building a semantically driven BDI clustering mechanism based on the output lightweight semantic features and semantic consistency matrix; performing message-passing neural network cooperative inference and parameter updates based on the cooperative topology; iteratively selecting the optimal cooperative node set using a semantic utility-driven long-distance meta-path progressive search strategy; and performing MPNN cooperative inference and model closed-loop updates based on the optimal cooperative node set, ultimately outputting the final position estimate of the non-cooperative radiation source target. This invention solves the problems faced by existing large-scale UAV swarms in non-cooperative radiation source localization, such as difficulties in signal aliasing disambiguation, low cooperative efficiency under non-independent and identically distributed data, and inaccurate node selection strategies.

A method for generating meeting minutes includes the follow steps. A video signal, an audio signal and source localization information of a video conference are obtained. Face recognition is performed on multiple image frames of the video signal to obtain multiple face recognition results. Voice recognition is performed on multiple audio segments of the audio signal to obtain multiple voice recognition results at multiple timestamps. The voice recognition results are matched with the face recognition results according to the source localization information, in order to obtain multiple speaker's identities. Speech to text transcription is performed on the audio segments of the audio signal to obtain a transcript. The speaker's identities are attached to the transcript according to the timestamps, in order to obtain a context. Context understanding is performed on the context to obtain a meeting minutes report.

This invention relates to an early warning system and method for ground pressure disasters in deep metal mining. The system combines high-sensitivity microseismic sensors, distributed acquisition units, signal processing and recognition technology, three-dimensional wave velocity model construction, data visualization, and a multi-parameter fusion early warning algorithm to achieve real-time, accurate monitoring and intelligent early warning of the entire rock mass fracturing process. The system hardware layer includes a 14Hz Geophones sensor network deployed in different deep sections, a netADC / netSP acquisition and processing module, and a UPS power supply system. The software layer includes modules for seismic source localization, feature extraction, deep learning classification, multi-parameter fusion model construction, and dynamic early warning output, supporting multi-platform operation and remote control. By constructing a three-dimensional wave velocity model, identifying microseismic signals in real time, and extracting multiple parameters such as energy, frequency, cumulative apparent volume, and Schmidt number, the system can achieve an early warning accuracy rate of ≥90%, significantly improving the ability to prevent and control ground pressure disasters in mines under complex geological conditions.

This invention provides a distributed radar noncoherent direct localization method, device, terminal, and medium for data anomaly scenarios, relating to the field of radar technology. The method includes: acquiring echo signal data received by each observation array; constructing a signal model characterizing the mapping relationship between the echo signal data and potential source locations, where the common detection area of ​​each observation array is a two-dimensional grid, the grid point coordinates correspond to the potential source locations, the signal model does not contain phase synchronization parameters between observation arrays, and is adapted to the noncoherent characteristics of the signal data; and constructing a system based on the signal model... 2,1 This invention addresses the group sparse optimization problem of norms. A pre-defined optimization algorithm is used to solve the problem, yielding a row sparse matrix. Based on this matrix, the true source localization result is determined, and the column vector elements of the row sparse matrix correspond to the potential source reflection energy estimation results at the grid points. This invention utilizes echo signal data from various observation arrays to achieve primary localization, eliminating secondary localization errors and reducing abnormal noise interference.

The application discloses a multi-sound source positioning and imaging method based on sparse representation and variational Bayesian inference, and steps are as follows: (1) an initial sound intensity matrix is generated under low resolution by using a conventional beamforming algorithm to estimate the sound source position; (2) the signal intensity gradient is calculated, the search position is updated along the gradient direction, and the accurate sound source position and high-resolution sound intensity matrix are obtained; (3) a sparse dictionary learning algorithm is used for sparse coding and dictionary update optimization of the high-resolution sound intensity matrix; (4) a variational Bayesian inference model is constructed based on the sparse coefficient, the lower bound of variation is optimized, and the posterior positioning estimation of the multi-sound source is carried out; and (5) the positioning result is fused with the camera image to obtain the visualized position of the sound source. The method realizes high-precision real-time positioning and imaging in a complex sound field by combining low-resolution positioning, gradient optimization, sparse dictionary learning and Bayesian inference, and has high spatial resolution and strong anti-interference capability.

The present application discloses systems and methods for developing acoustic and touch interfaces using one or more structural vibration sensors affixed to a surface. The method utilizes the resonant properties of the structure and machine learning to infer information about the source such as the position of a sound source in a room or the location the structure was touched. The application further discloses systems utilizing methods may reduce the number of sensors needed for applications such as sound-source localization, acoustic beamforming and touch interfacing, reduce the manufacturing cost of implementing the systems, and improve device durability when compared with systems currently used for the aforementioned applications.