Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

191results about How to "Guaranteed Computational Efficiency" patented technology

Method and system for detecting false data injection attack of power system

The invention discloses a power system false data injection attack detection method and system. Measurement data of a power distribution system are acquired; constructing an attack sequence, and injecting the constructed attack sequence into the obtained measurement data to simulate a false data attack; constructing a false data injection attack detection model based on the expansion causal convolution and the recurrent neural network to analyze the state change of the power distribution system; and judging an attack behavior according to the analyzed state of the power distribution system. According to the method, time sequence data can be processed, the long-term dependency relationship and the complex mode can be captured, and the response capability to the complex attack mode can be improved, so that the false data injection attack can be effectively detected and prevented, and the safety and reliability of the power distribution system can be maintained.
Owner:STATE GRID LIAONING ELECTRIC POWER CO LTD +1

Wave water power numerical simulation method for sea area provided with breakwater

The invention, which relates to the technical field of offshore wave simulation, discloses a wave hydrodynamic numerical simulation method for a sea area with a breakwater, comprising the following steps: S1, determining a target sea area, and collecting sea chart data and water depth topographic data of the target sea area; s2, determining a shoreline position of the target sea area, and establishing a water depth terrain database of the target sea area; s3, setting grid precision, and constructing a wave hydrodynamic two-dimensional model of the target sea area; s4, setting boundary conditions for driving the wave hydrodynamic two-dimensional model, and setting initial parameters; and S5, based on the wave hydrodynamic two-dimensional model constructed in the S3, performing numerical simulation on the target sea area by using a sub-grid method to obtain the characteristic wave height and periodic data of the target sea area. According to the method provided by the invention, the calculation efficiency can be ensured, the wave transmission characteristics of the open type oyster reef are considered, and the influence of open type oyster reef construction on the wave hydrodynamic environment field can be accurately simulated.
Owner:OCEAN UNIV OF CHINA

Scene reconstruction Gaussian model generation method and scene reconstruction method

The invention discloses a scene reconstruction Gaussian model generation method and a scene reconstruction method. Comprising the following steps: determining training data according to a to-be-trained image sequence, training a Gaussian model according to the training data, and generating a scene reconstruction Gaussian model; determining training data according to the to-be-trained image sequence: determining point cloud data according to the to-be-trained image sequence and constructing anchor Gaussian sets of different levels; partitioning the to-be-trained image sequence, determining an image acquisition device corresponding to each image partition, projecting anchor points in different layers of anchor Gaussian sets, determining all observable anchor points of each image acquisition device, and generating a training set of each image partition; determining a constraint expression of each to-be-trained image in the training image sequence according to the to-be-trained image sequence and the three-dimensional Gaussian parameters; and the training set and the constraint expression of the image partition are used as training data, so that the problem of poor scene reconstruction quality is solved, the calculation efficiency and the reconstruction precision are ensured, and the reconstruction quality is improved.
Owner:SHANGHAI ARTIFICIAL INTELLIGENCE INNOVATION CENT

Wire instance detection method and system based on key point matching

The invention discloses a conductor instance detection method and system based on key point matching, and the method comprises the steps: obtaining original aerial image data, generating a four-dimensional feature matrix through dynamic feature mapping and enhancement processing, obtaining three-dimensional projection feature data through projection transformation, and carrying out the feature enhancement to obtain enhanced feature data. Generating enhanced feature map data through multi-scale feature fusion; performing feature recombination on the enhanced feature map data to obtain feature matrix data, generating response matrix data through recursive decomposition and mapping, and calculating adaptive threshold data to screen key points; generating thermodynamic diagram data for the screened key point set through a multi-layer thermodynamic diagram, obtaining a spatial incidence matrix by combining spatial feature enhancement, and performing endpoint matching based on multi-dimensional constraints; and finally, through boundary feature analysis and cross-block instance matching, a global traverse map is constructed, and accurate detection of traverse instances is realized. According to the invention, the accuracy and robustness of wire detection can be effectively improved.
Owner:ZHONGKE FANGCUN ZHIWEI (NANJING) TECH CO LTD

Low-light image denoising method based on local-global interactive Transform

The invention discloses a low-light image denoising method based on a local-global interactive Transform, and the method is characterized in that the method comprises the following steps: 1, carrying out the denoising of a low-light image; the method comprises the following steps: step 1, constructing an overall structure of LGIT for recovering a given degraded image; 2, a Top-N Patches selection module is used for selecting N patches most related to the target patch and providing feature input for subsequent modules; step 3, the Patch selects an attention Transform module, and the Transform module is used for carrying out global feature aggregation based on the information of Top-N Patches; and step 4, mixing the double-gate feed-forward network MDGFF for further processing the features output by the PSATB and generating a final de-noised image. According to the method, the self-attention inside the patch and between the patches is effectively calculated on different scales, and accurate removal and detail recovery of low-light image noise are realized.
Owner:ZHONGBEI UNIV

Container factory container truck path planning method and system based on Beidou positioning

PendingCN119063716AAchieve precise digital expressionAccurately reflect real-time traffic conditionsInstruments for road network navigationInternal combustion piston enginesRoad networksTransportation technology
The invention provides a container factory container truck path planning method and system based on Beidou positioning, and relates to the technical field of traffic transportation, and the method comprises the steps: collecting geographic space data, dynamic operation data and container truck transportation data, the container truck transportation data comprises real-time position information and real-time state information of a container truck; constructing a live-action three-dimensional model corresponding to the geographic space data based on the geographic space data; extracting road network information from the live-action three-dimensional model to identify each transportation node in a transportation area and a link formed between any two adjacent transportation nodes; constructing a dynamic traffic distribution model according to the real-time position information and the real-time state information; inputting the container truck average density, the container truck average speed and the container truck average flow in each link into the dynamic traffic distribution model to predict state variables of all container trucks in a transportation area in the container factory; and based on the state variables and the dynamic operation data, generating an optimal transportation path for each container truck in combination with a target path optimization function and a time cost function. According to the invention, the accuracy of container truck scheduling execution can be improved.
Owner:WUHAN YUHUIHONG TECH CO LTD

Automatic three-dimensional building reconstruction method

The invention relates to an automatic three-dimensional reconstruction method for a building. The invention relates to the technical field of civil engineering infrastructure health monitoring, and the method comprises the steps: controlling an unmanned aerial vehicle to fly around a building and shoot building facades at different visual angles, and obtaining a multi-visual-angle picture set; based on an SfM algorithm, calculating camera parameters of each picture and obtaining an initial point cloud; according to the obtained camera parameters, calculating a space boundary of the building point cloud and generating a pseudo-view camera; according to the distance and direction weight of the camera, defining and optimizing the Gaussian ellipsoid attribute of an unknown view angle by the Gaussian ellipsoid of the adjacent known view angle; and filtering the building point cloud optimization Gaussian model based on a density DBSCAN algorithm to obtain a high-precision building three-dimensional model. The method is convenient and fast, does not depend on manual parameter adjustment, has low requirement on the overlapping rate of the pictures shot by the unmanned aerial vehicle, is not influenced by illumination, and improves the reconstruction efficiency of the building three-dimensional model and the precision of the geometric detail area.
Owner:HARBIN INST OF TECH

Medical image target detection method and device, electronic equipment and readable storage medium

The invention relates to a medical image target detection method and device, electronic equipment and a readable storage medium, and the method comprises the steps: obtaining a to-be-detected medical image and a medical image target detection model, according to the medical image target detection model, an expandable residual module is added in a backbone network, a multi-branch attention mechanism module is added to a head in a neck network, and at least part of conv modules in the backbone network and the neck network are replaced by YOLOv8 models of deep down-sampling modules. The medical image target detection model is obtained through joint regression loss function training; and inputting the to-be-detected medical image into the medical image target detection model to obtain a medical image target detection result. According to the medical image target detection method provided by the invention, the accuracy of medical image target detection is improved.
Owner:ANHUI UNIV OF SCI & TECH

Inter-vehicle relevance traffic monitoring method based on space-time analysis and track matching

The invention belongs to the technical field of traffic control, and relates to an inter-vehicle relevance traffic monitoring method based on space-time analysis and track matching, which comprises the following steps: firstly, collecting vehicle structured space-time data, and defining a space-time data track and a space-time data point; detecting vehicle trajectory features to obtain a trajectory similarity score; detecting vehicle parking points to obtain a similarity score of the parking points; and then adding the trajectory similarity score and the stay point similarity score to obtain a relevance score between the vehicles, not only performing trajectory similarity analysis in combination with an existing deep learning technology, but also introducing a stay point detection and matching technology, comprehensively analyzing the relevance between the vehicles from two dimensions of a spatio-temporal data trajectory and a spatio-temporal data point, and obtaining the relevance between the vehicles. The potential relevance between the vehicles can be captured and analyzed more comprehensively and accurately, and the accuracy and practicability of the monitoring system are improved.
Owner:QINGDAO SONLI SOFTWARE INFORMATION TECH +1

Robust probability hypothesis density SLAM method under heavy tail noise

The invention relates to the technical field of simultaneous localization and mapping, in particular to a robust probability hypothesis density (SLAM) method under heavy-tailed noise, which is characterized in that an SLAM problem is decomposed into two coupled sub-problems of robot pose estimation and map estimation, robot pose modeling is particle modeling, environment map modeling is probability hypothesis density, and the robot pose estimation is particle modeling; secondly, carrying out importance sampling on all particles representing the pose of the robot by utilizing an EKF (Extended Kalman Filter) algorithm; performing map prediction based on each particle, and adding the new probability hypothesis density into the posterior probability hypothesis density; modeling the measurement noise into student t distribution, and deducing a complete map updating expression by using an edge approximation likelihood function based on the student t distribution; and finally, updating the weight of the pose particles of the robot by using an SC strategy, and carrying out state extraction and particle resampling. According to the method, the estimation precision of the SLAM problem in the heavy-tailed noise scene is effectively improved.
Owner:GUILIN UNIV OF ELECTRONIC TECH +1

Three-dimensional back projection imaging method and device of array interference SAR (Synthetic Aperture Radar)

The invention discloses a three-dimensional back projection imaging method and a three-dimensional back projection imaging device for an array interference SAR (Synthetic Aperture Radar), which belong to the field of SAR three-dimensional imaging, and comprise the following steps: replacing registration by utilizing a three-dimensional grid interpolation method based on a low-altitude scene three-dimensional array SAR imaging model of a polar coordinate system and a Cartesian coordinate system, expanding a sensing matrix, and utilizing compressed sensing to realize the three-dimensional back projection imaging of the array interference SAR. And the three-dimensional information is reconstructed by a CS algorithm, and high-quality three-dimensional reconstruction of an observation scene is accurately realized. According to the method, on the premise of ensuring the calculation efficiency, the resolution and accuracy of the three-dimensional imaging of the array interference SAR under the low-altitude condition are improved, the reconstruction quality of the overlay area is remarkably improved, and the three-dimensional entropy index in the measured data is improved by about 0.1.
Owner:AEROSPACE INFORMATION RES INST CAS

Human body behavior recognition method and system based on space-time decoupling multi-scale feature fusion

The invention discloses a human body behavior recognition method and system based on space-time decoupling multi-scale feature fusion. The system comprises a local space-time multi-scale feature fusion network MDCNN and a space-time decoupling network DTST. Extracting different scale features of multichannel sensor data by using a multi-scale convolutional neural network, extracting shallow layer features and deep layer features of signals at the same time, and extracting features representing action information by fully utilizing complementary information among different level features; the method comprises the following steps of: decoupling a network into a time decoupling structure and a space decoupling structure through a time-space decoupling network based on Transform, so as to accurately extract time-space characteristics in an inertial signal; and finally, using a Softmax classifier to carry out action classification and recognition. According to the method, consistency and complementarity among multi-sensor data can be efficiently utilized, multi-scale spatial-temporal characteristics are fused, changes in time dimensions and relevant information in spatial positions are fully considered, different attention degrees are kept for signals of different sensing channels, and the recognition precision in human behavior recognition is improved.
Owner:XIAN UNIV OF TECH +1

Method for calculating blocking effect of towing tank of underwater vehicle

The invention relates to the technical field of computational fluid mechanics, in particular to a method for calculating the blocking effect of an underwater vehicle towing tank, and the method comprises the steps: building an underwater vehicle physical model; defining each boundary condition of the CFD numerical calculation domain, and solving the total resistance and each resistance component of the underwater vehicle based on the calculation domain grid and the configured numerical calculation condition; and post-processing the total resistance and each resistance component to obtain a calculation numerical result of the underwater vehicle under different blocking ratio working conditions. According to the method, by verifying the independence of the grids, the appropriate number of the grids is determined, and the calculation precision and efficiency are guaranteed.
Owner:NORTHWESTERN POLYTECHNICAL UNIV

Complex terrain wind field refined simulation method based on high-resolution grid

PendingCN119647340AImprove refined simulation capabilitiesfast splitDesign optimisation/simulationCAD numerical modellingTerrainWind engineering
The invention relates to the fields of wind engineering, weather forecast, environmental protection and aviation, and provides a complex terrain wind field refined simulation method based on a high-resolution grid. The objective of the invention is to solve the problem of simulation errors caused by different topographic data formats and coordinate systems in complex topographic wind field simulation and to improve simulation precision and efficiency. According to the main scheme, the method comprises the following steps: generating a three-dimensional computational grid of wind mine software according to satellite image elevation data, and obtaining a three-dimensional unstructured grid for wind field CFD simulation; selecting a proper entrance boundary condition type for four side boundary surfaces and a top surface of the target area; and wind mine software is driven to carry out numerical calculation to obtain wind field information data of the refined simulation of the complex terrain. The method is used for wind field characteristic evaluation under complex terrains, and is beneficial to efficient utilization of wind energy, weather forecast, environmental protection and aviation safety.
Owner:SICHUAN GREEN ENERGY INTELLIGENT COMPUTING TECHNOLOGY CO LTD

NeRF modeling method fusing time dimension and hierarchical sampling strategy

The invention discloses a NeRF modeling method fusing a time dimension and a hierarchical sampling strategy. The NeRF modeling method comprises the following steps: constructing a data set; sampling M coarse sampling points for each ray of the RGB image in the data set; position coding is carried out on the coarse sampling points, and then the volume density and color of the coarse sampling points are output by the MLP; calculating the weight of the coarse sampling point, and accumulating the weight; n fine sampling points are generated based on the accumulated weight, and the volume density, the color, the opacity and the accumulated transmittance of the fine sampling points are calculated; calculating the color of the ray by combining the coarse sampling point and the fine sampling point; constructing a loss function L of the MLP; and training the MLP to obtain a three-dimensional scene modeling model. According to the method, the rendering efficiency is improved through a hierarchical sampling strategy, an explicit space structure is fused during coding to enhance the capturing capability of the model for complex geometry, and time coding is added to efficiently cope with dynamic changes in a scene, so that better performance and efficiency are achieved when large-scale, complex and dynamic three-dimensional scenes are processed.
Owner:SHENZHEN SENSING DATA TECH CO LTD +1

Traffic small target detection method, device and equipment for high-speed driving vehicle and medium

The invention relates to a traffic small target detection method and device for a high-speed driving vehicle, equipment and a medium, and relates to the field of image processing, and the method comprises the steps: inputting a traffic scene image frame into a second traffic small target detection model which is trained to a convergence state, and carrying out the detection of a traffic scene image frame in a plurality of multi-layer stacked recursive Star Block modules of a StarNet lightweight network, performing element-by-element multiplication on image features corresponding to the traffic scene image frames by adopting star operation to exponentially increase implicit feature dimensions in the traffic scene image frames so as to determine high-dimensional image features of the traffic scene image frames; and inputting the high-dimensional image features into a detection head network in a traffic small target detection model to obtain a plurality of traffic small targets in the traffic scene image frame so as to complete traffic small target detection of the high-speed driving vehicle. According to the method, the small traffic target can be accurately and quickly detected in a traffic scene in high-speed driving, and higher target detection performance can be realized in a resource-limited environment such as high-speed driving vehicles.
Owner:SOUTH CHINA UNIV OF TECH

Markov chain Monte Carlo-based paleoclimate inversion method, system and equipment

The invention belongs to the technical field of climate prediction, particularly relates to a Markov chain Monte Carlo-based paleoclimate inversion method, system and equipment, and aims to solve the problems of accuracy and high efficiency of tracking climate change details in global inversion. The method comprises the following steps: determining a to-be-inverted paleoclimate index; inputting the to-be-inverted paleoclimate index into a paleoclimate inversion model to obtain a paleoclimate inversion result output by the paleoclimate inversion model; wherein the paleoclimate inversion model is obtained by performing iterative matching based on a target curve of a paleoclimate index and an initial parameter set determined based on the target curve of the paleoclimate index so as to continuously update a to-be-inverted parameter set. According to the method, in global inversion, climate change details are tracked more accurately, calculation is more efficient, and the response capability and adaptability to the time sequence are improved.
Owner:INSTITUTE OF GEOLOGY AND GEOPHYSICS CHINESE ACADEMY OF SCIENCES

Wave water power numerical simulation method for sea area provided with hedge shrub dam

The invention discloses a wave hydrodynamic numerical simulation method for a sea area provided with a hedge shrub dam, and relates to the technical field of offshore wave simulation, and the method comprises the following steps: S1, determining a target sea area, and collecting actually measured water depth topographic data and chart water depth data of the target sea area, the target sea area being a salt marsh beach sea area where the hedge shrub dam is planned to be arranged; s2, setting grid precision, and constructing a wave hydrodynamic two-dimensional model of the target sea area by using a structured grid; s3, setting boundary conditions for driving the wave hydrodynamic two-dimensional model, and setting initial parameters; and S4, based on the wave hydrodynamic two-dimensional model constructed in the S2, performing numerical simulation on the target sea area by using a sub-grid method to obtain the characteristic wave height and periodic space-time distribution of the target sea area. According to the method, the calculation efficiency can be guaranteed, the wave transmission characteristics of the hedge shrub dam are considered, and the influence of construction of the hedge shrub dam on the wave hydrodynamic environment field can be accurately simulated.
Owner:OCEAN UNIV OF CHINA

Large 3D scene texture dynamic loading and unloading method, medium and system

The invention provides a large 3D scene texture dynamic loading and unloading method, medium and system, and belongs to the technical field of 3D scene loading.The large 3D scene texture dynamic loading and unloading method, medium and system comprise the steps that three-dimensional scene parameters are collected, texture frequency parameters are decomposed to obtain basic and fluctuation components, a texture loading model is constructed, and a coefficient matrix is generated; packaging the texture file into a fixed-size block, building a double-layer resource optimization model based on a coefficient matrix, setting global and local optimization targets, and calculating a comprehensive contribution coefficient; an initial resource allocation scheme is obtained through an evolutionary algorithm, and a final scheduling scheme is obtained through dynamic programming; according to the texture blocks, loading the texture blocks, determining a recovery strategy in combination with life cycle, grouping association and state transition prediction, and executing resource recovery; the process comprises texture analysis, model construction, resource optimization, scheduling and recovery, and efficient management of three-dimensional scene textures is realized; the resource scheduling decision-making mechanism can be optimized, and the user experience is improved.
Owner:SHANDONG GUANXIANG WORLD DIGITAL TECH CO LTD

Neural network target detection method based on FPGA acceleration

The invention discloses a neural network target detection method based on FPGA acceleration, and relates to the technical field of convolutional neural network accelerators. In order to solve the technical problems that in the design of an existing FPGA-based convolutional neural network accelerator in the prior art, the accuracy of a YOLO network is remarkably reduced, and the delay is high, the technical scheme provided by the invention is as follows: the method for constructing the neural network target detection system based on FPGA acceleration comprises the following steps: obtaining a compressed network model; carrying out data processing and transmission to obtain a preprocessed feature map and weight data; performing convolution calculation on the preprocessed feature map and the weight data to obtain a convolution calculation result; a convolution calculation result is processed through a maximum pooling layer, an Upsample up-sampling layer and a Concat layer, and a pooling, up-sampling and splicing feature map is obtained; and the NMS non-maximum suppression module containing the FIFO structure performs post-processing on the pooled, up-sampled and spliced feature map. The method is suitable for the work of neural network target detection.
Owner:HARBIN INST OF TECH

AI auxiliary radiotherapy planning method based on diffusion model

The invention discloses an AI auxiliary radiotherapy planning method based on a diffusion model, and belongs to the technical field of artificial intelligence. The method comprises the following steps: S1, collecting and processing a data set: collecting a positioning CT image, sketching information and clinical information of a cancer patient, and establishing a cancer radiotherapy dose database; s2, a dose prediction model: establishing a deep learning network based on a deep diffusion generation model based on task and data characteristics, and obtaining a model with higher generalization ability by using the characteristics of strong generation ability, high result authenticity and rich details of the deep learning network; and S3, in combination with a 2.5 D input strategy of individualized information, a 2.5 D mode is adopted, an efficient and lightweight high-precision model is realized, and meanwhile, the model established in the mode is more easily applied to clinic. According to the method, the AI technology is introduced, the slicing strategy of 2.5 DCT data is combined, and the structure and performance of the model are optimized by combining the two-dimensional and three-dimensional characteristics.
Owner:RIZHAO PEOPLES HOSPITAL OF SHANDONG PROVINCE

Power transmission line and tower damage early warning method and device under strong wind effect

The invention discloses a damage early warning method and device for a power transmission line and a tower under the action of strong wind. The temperature and wind characteristic data of the power transmission line and each preset part of the power transmission tower are collected in real time through a plurality of sensors; acquiring image information of the power transmission line and the preset part of the power transmission tower in real time through a plurality of image acquisition devices; obtaining a first stress distribution result of each preset part according to the currently obtained temperature and wind characteristic data of each preset part in sequence; and according to the current image information of each preset part and the first stress distribution result, obtaining a damage early warning result of the power transmission line and the power transmission tower at the current moment. Through the method and the device, the stress distribution result of each preset part can be obtained, so that risk points can be quickly checked, in addition, the stability of the tower and the power transmission line is accurately evaluated by considering the real-time building material change condition of each preset part in the tower and the power transmission line, and the damage early warning accuracy is improved.
Owner:GUANGDONG POWER GRID CO LTD

Moving object self and multi-target pose multi-sensor fusion simultaneous estimation method oriented to meta-universe environment

The invention belongs to the field of meta-universe object pose estimation technology and robots, and discloses a meta-universe environment-oriented moving object self and multi-target pose multi-sensor fusion simultaneous estimation method. A more robust object-level data association is established using a multi-probability fusion tracker based on a Kalman filter and points in the optical flow. On the basis, a dynamic point cloud elimination method fusing the speed prior of a 3D Kalman filter and camera optical flow decoupling is provided so as to improve the precision of self pose estimation of the speedometer. And finally, combining multiple constraint factors in the proposed framework to refine the poses of the object and the object. Evaluation is carried out in a KITTI original data set, and compared with a reference and most advanced method, the method obtains better attitude precision of the KITTI and a tracking target. In addition, the proposed system shows the feasibility of real-time performance so as to ensure the practical application value of the system.
Owner:NORTHEASTERN UNIV CHINA +1

Intelligent individual soldier system based on AI multi-mode

The invention relates to the field of intelligent individual soldiers, in particular to an intelligent individual soldier system based on AI multiple modes. The method comprises the following steps: acquiring and preprocessing multi-modal original data, extracting initial features from the multi-modal preprocessed data, and carrying out standardization processing to obtain multi-modal standard features; the multi-modal standard features are optimized, and a comprehensive feature matrix is obtained; analyzing the comprehensive feature matrix, and extracting spatial features; inputting the spatial features into a gating circulation unit to obtain a time sequence state, and predicting the time sequence state to generate a prediction result; and the prediction result is transmitted to inspection personnel. The problem that although a traditional convolutional neural network and a long-short-term memory network can effectively process spatio-temporal data, long-time stable operation in a resource-limited single-soldier system is difficult due to high calculation complexity and resource consumption is solved. The problems that in the feature fusion process, fixed weights cannot dynamically adapt to environment changes, and a traditional feature processing method is low in robustness to noise and abnormal values are solved.
Owner:GUODIAN GANSU NEW ENERGY CO LTD

Personalized federal learning method and system

The invention discloses a personalized federal learning method and system, and the method comprises the steps: introducing an scSE attention mechanism at a client, adding the scSE attention mechanism into a generator and a discriminator of a GAN, obtaining important features of data in channel and spatial dimensions, and carrying out the enhancement of the features; by adaptively adjusting network parameters, optimizing a generative adversarial network structure, constructing a client local model, and uploading the client local model to a central server; and at the central server side, constructing a personalized federal learning model, receiving the local model uploaded by the client side, carrying out aggregation training on the local model, obtaining an optimized global model, and sending parameters of the optimized global model to each client side. Key features in the data are adaptively concerned, the training efficiency and the data generation quality of the GAN in personalized federated learning are improved, the performance of the model on various tasks is improved, the method adapts to the differences of data sets on distribution and characteristics, and the problem of heterogeneous data is solved.
Owner:SHANGHAI UNIVERSITY OF ELECTRIC POWER

Intelligent individual soldier system based on AI multi-mode

The invention relates to the field of intelligent individual soldiers, in particular to an intelligent individual soldier system based on AI multiple modes. The method comprises the following steps: acquiring and preprocessing multi-modal original data, extracting initial features from the multi-modal preprocessed data, and carrying out standardization processing to obtain multi-modal standard features; the multi-modal standard features are optimized, and a comprehensive feature matrix is obtained; analyzing the comprehensive feature matrix, and extracting spatial features; inputting the spatial features into a gating circulation unit to obtain a time sequence state, and predicting the time sequence state to generate a prediction result; and the prediction result is transmitted to inspection personnel. The problem that although a traditional convolutional neural network and a long-short-term memory network can effectively process spatio-temporal data, long-time stable operation in a resource-limited single-soldier system is difficult due to high calculation complexity and resource consumption is solved. The problems that in the feature fusion process, fixed weights cannot dynamically adapt to environment changes, and a traditional feature processing method is low in robustness to noise and abnormal values are solved.
Owner:GUODIAN GANSU NEW ENERGY CO LTD

Power grid optimal power flow calculation method based on convex relaxation technology and related equipment

The embodiment of the invention discloses a power grid optimal power flow calculation method based on a convex relaxation technology and related equipment, first active power flow, first reactive power flow, second active power flow and second reactive power flow can simulate operation of alternating current and direct current power grids including high-voltage direct current power transmission, and through the convex relaxation technology, the optimal power flow of the power grid can be calculated. And a non-convex optimization problem in the power flow is converted into a convex optimization problem which is easy to solve, so that the accuracy and the calculation efficiency of the optimal power flow are ensured by calculating the optimal power flow by using the branch power flow model optimized by the convex relaxation technology, and the analysis requirement of an alternating current and direct current hybrid power grid is met.
Owner:YUNNAN POWER GRID CO LTD ELECTRIC POWER RES INST

Semi-active suspension control system based on Transform architecture

The invention belongs to the technical field of vehicle suspension system control, and discloses a semi-active suspension control system based on Transform architecture, which comprises the following steps: arranging a sensor network, and collecting vehicle dynamic data; using a preset algorithm to align asynchronous data from different sensors, and keeping time sequence consistency of heterogeneous data; the measured value of the sensor i at the time t is xi (t), in the invention, the weight distribution of the state quantity is reduced through a Transform mechanism, the error control quantity caused by the sensor error is avoided, the robustness of the suspension control system to the data exception is improved, and the reliability of the suspension control system is improved. According to the method, stable control can be kept under sensor faults, external attacks and false sensor data injection or extreme road conditions, through a DDPG reinforcement learning control algorithm, the adaptive capacity of the system in a complex and changeable environment is enhanced, different road conditions, loads and driving styles can be rapidly adapted, and the control effect is improved while the calculation efficiency is guaranteed.
Owner:JILIN UNIVERSITY

Mountain area bridge dynamic response analysis method under transverse bridge direction earthquake and landslide coupling effect

The invention provides a mountainous area bridge dynamic response analysis method under the transverse bridge direction earthquake and landslide coupling effect, and aims to analyze the bridge dynamic response under the transverse bridge direction earthquake and landslide coupling effect and reveal the bridge dynamic response and damage mechanism under the coupling effect. The method comprises two core parts: 1, a landslide dynamic impact effect simulation method: based on a discrete element analysis method, defining a discrete element contact model, establishing a mountain-pier overall model, generating landslide body particles and recording a landslide effect time history; and 2, a transverse bridge direction earthquake and landslide coupling effect simulation method: based on a finite element analysis method, establishing a bridge monomer finite element model, determining a coupling effect moment, and carrying out bridge dynamic response analysis. According to the method, a step-by-step numerical model of integrating first and then single bodies is adopted, the calculation precision and efficiency are guaranteed, and the method can provide scientific support for earthquake disaster prevention and risk assessment of mountainous area bridges.
Owner:FUZHOU UNIV

Vehicle adaptive model prediction control method

The invention relates to a vehicle adaptive model prediction control method. The method comprises the following steps: collecting vehicle dynamic parameters, and constructing a two-degree-of-freedom vehicle dynamic model according to the dynamic parameters; constructing a first matrix used for describing an input and output relation of the linear state space equation; determining a model order of a vehicle model corresponding to the two-degree-of-freedom vehicle dynamics model through an akaike information criterion, and identifying parameters of the first matrix through a recursive least square method of introducing a forgetting factor to obtain the vehicle model; self-adaptively adjusting forgetting factors in the vehicle model through fuzzy logic and identification residual errors to obtain an optimized vehicle model; and establishing a vehicle controller based on the optimized vehicle model to control a vehicle control system in real time. According to the method, effective tracking of the expected state of the vehicle under different working conditions can be realized, the identification speed and the modeling precision are improved, and the modeling complexity is reduced.
Owner:JILIN UNIVERSITY