Patents
Literature
70results about How to "Reduce learning difficulty" patented technology
Efficacy Topic
Property
Owner
Technical Advancement
Application Domain
Technology Topic
Technology Field Word
Patent Country/Region
Patent Type
Patent Status
Application Year
Inventor
Integrated CAD drawing set drawing template system and method thereof
The invention relates to the technical field of building design, and particularly discloses an integrated CAD drawing set drawing template system and method, and the system comprises a drawing management module, a template presetting module, a graph reference module, an attribute configuration module, a workflow management module, and a quality inspection module. The drawing management module is used for storing and maintaining all drawings and metadata associated with the drawings and realizing automatic naming and classification of the drawings; the template presetting module is used for generating a template file meeting requirements according to a standard set by a user; the graph reference module is used for automatically introducing related graph files into the layout by using a CAD graph reference system; the attribute configuration module is used for automatically adjusting character information and a layer display state of a drawing according to a drawing environment and an attribute standard input by a user; through the integrated design, the convenient operation process and the abundant preset templates, the problems that a traditional drawing set is high in learning cost and complex in operation are successfully solved, and the drawing efficiency and quality are greatly improved.
Owner:ZHENGZHOU QUANXIANG MARKING TECHNOLOGY CO LTD
Fighter flight control method based on adversarial skill embedding and hierarchical reinforcement learning
ActiveCN118759921BReduce study loadReduce learning difficultyProgramme controlComputer controlSimulationReinforcement learning algorithm
The present invention discloses a fighter flight control method based on adversarial skill embedding and hierarchical reinforcement learning. The method designs an adversarial skill embedding learning framework composed of an adversarial generative algorithm and a deep reinforcement learning algorithm. The method uses expert flight data to automatically learn the mapping of underlying control instructions to macro maneuvering skills by an AI algorithm. By reusing these macro maneuvering skills, the method significantly reduces the learning difficulty of downstream air combat tasks, greatly compresses the convergence time of the reinforcement learning algorithm, improves the game performance of air combat intelligent agents, and gives air combat intelligent agents higher flexibility and rapid migration capabilities, thereby exploring a new technical path for the practical training and deployment of air combat intelligent agents.
Owner:ZHEJIANG LAB
Photovoltaic power prediction method based on improved empirical mode decomposition and optimized long short-term memory network
PendingCN121863356ARealize global optimizationimprove accuracyGeneration forecast in ac networkPhotovoltaic monitoringOutlier eliminationPredictive methods
The invention discloses a photovoltaic power prediction method based on improved empirical mode decomposition and an optimized long short-term memory network, and the method comprises the steps: firstly carrying out the preprocessing of abnormal value elimination, missing value filling, normalization and the like of photovoltaic power and related meteorological data, and improving the data quality; then, an improved empirical mode decomposition (EE-ANEMD) algorithm is adopted to decompose the preprocessed power sequence into a multi-scale intrinsic mode function component and a residual term, and high-frequency noise, intermediate-frequency fluctuation and a low-frequency trend are effectively separated; global optimization is carried out on the hidden layer unit number, the initial learning rate and the maximum number of training times of the LSTM network through an improved sparrow search algorithm (ISSA), finally, the optimized LSTM is utilized to carry out training prediction on each component, and results are fused and subjected to reverse normalization to obtain a final value. Experiments show that the test set RMSE of the method is reduced compared with that of a single LSTM, the mid-term prediction precision is remarkably improved, and reliable technical support is provided for power system dispatching, new energy consumption planning and photovoltaic power station operation and maintenance.
Owner:HUZHOU ELECTRIC POWER SUPPLY CO OF STATE GRID ZHEJIANG ELECTRIC POWER CO LTD +3
Art pigment viscosity testing device
InactiveCN119470154AReduce learning difficultyFast testDirect flow property measurementViscosityPhysics
The invention relates to the technical field of viscosity testing, in particular to an art paint viscosity testing device which solves the problems that when a worker tests the viscosity of art paint, operation is tedious, and the accuracy of a testing result is poor. The art paint viscosity testing device comprises a circular shell, a liquid storage barrel is arranged in the circular shell, and the rear end of the liquid storage barrel is rotationally connected with a pressing plate and a baffle; a sliding block is slidably connected to the rear end of the liquid storage barrel and hinged to the rear end of the pressing plate through a pressing rod, a transparent plate is hinged to the front end of the circular shell, and a timer is arranged at the upper end of the circular shell. A waste liquid barrel is detachably connected to the lower portion of the round shell, a rotary disc is rotationally connected into the round shell through a vertical shaft, a scraper is arranged in the round shell, an overflow barrel and a quantifying cup are arranged on the rear portion of the rotary disc, a U-shaped plate is arranged at the lower end of the quantifying cup, a ball is rotationally connected to the vertical portion of the U-shaped plate, and a round hole is formed in the ball. According to the invention, automatic pigment adding and pigment outflow time testing can be realized, the testing is simpler and more convenient, influence factors can be reduced, and the accuracy of a testing result is improved.
Owner:WALL PAINTING (SHANDONG) SCULPTURE ART CO LTD
Retired power battery digital lossless recombination method of reconfigurable battery network
PendingCN119231691Adelay agingReduce charge and discharge timesBatteries circuit arrangementsCells structural combinationElectrical batteryNetwork topology
The invention relates to the technical field of battery reconfiguration of an energy storage battery system, in particular to a digital lossless reconfiguration method for retired power batteries of a reconfigurable battery network. The invention discloses a decommissioned power battery digital lossless recombination method of a reconfigurable battery network. The method comprises the following steps: S100, designing a reconfigurable battery network topological structure; S110, establishing the reconfigurable battery network; s120, establishing an optimization control framework; s130, estimating a battery equivalent circuit model and a parameter state; s140, establishing a dynamic optimization model; s200, controlling the reconfigurable battery network; S210, establishing a control framework; s220, arranging an energy storage strategy based on AOE control configuration; and S230, performing AOE-based energy storage system power optimization distribution double-layer control. A decommissioned power battery digital lossless recombination method of a reconfigurable battery network has the advantages that an energy storage battery pack intermittently undertakes charging and discharging tasks, the charging and discharging times are effectively reduced, SOC balance between energy storage battery modules is achieved, overcharging and overdischarging of energy storage batteries are avoided, and battery aging is slowed down.
Owner:LBATTERYCLOUD CO LTD +1
Electric scissors capable of adjusting cutting angle of scissors
ActiveCN222485452Ulow costsmall structureCutting implementsCuttersStructural engineeringElectric control
The utility model discloses a pair of electric scissors capable of adjusting the cutting angle of the scissors, which comprises a main body, and the main body is provided with a fixed blade, a movable blade and a sensor for controlling the rotating angle of the movable blade, the movable blade has a closed end position relative to the fixed blade; the sensor comprises an end position sensor for detecting that the movable blade reaches the end position; the tail position sensor is movably installed on the main body and can adjust the position relative to the main body so as to change the tail position where the movable blade can move to and adjust the closing degree of the movable blade and the fixed blade. The utility model has the advantage of solving the problem that the closing degree is reduced after the blade is abraded at present. The shearing compensation of the abraded movable blade is realized with lower cost and fewer structures without modifying an electric control program or modifying the existing movable blade and fixed blade, the operation difficulty is greatly reduced, the tail position of the movable blade can be changed only by adjusting the relative position of the tail position sensor on the main body, and the shearing compensation of the abraded movable blade is realized. The adjustment is simpler and quicker.
Owner:ZHEJIANG YAT ELECTRICAL APPLIANCE CO LTD
Embedded graphical programming development method and device based on Internet of Things and medium
PendingCN119668585AImprove work efficiencyreduce errorsSoftware designVisual/graphical programmingProgramming languageGraphics
The invention discloses an embedded graphical programming development method and device based on the Internet of Things and a medium, and the method comprises the steps: creating a graphical programming system application containing a Blockly framework, adding a dependency library on the graphical programming system application, and carrying out the page layout and initialization configuration; the method comprises the following steps of: constructing graphic block components with different functions on the application of the graphic programming system, and determining required graphic blocks: a Modbus RTU (Remote Terminal Unit) graphic block component, a serial port-Rs485 graphic block component, a NetManage graphic block component, a Home Assistance graphic block component, a ThingsBoard graphic block component and an input / output control graphic block component; freely combining the determined required graphic blocks to build a business scene, and completing program development through graphic programming; and downloading the program subjected to graphical programming to equipment for running. The development efficiency of the system is improved, and the development difficulty is reduced.
Owner:BEIJING NEWLAND TIMES EDUCATION TECH CO LTD
Real-time advertisement evaluation decision-making system based on large language model agent
ActiveCN119313403AReduce the difficulty of adaptationReduce learning difficultyAdvertisementsText processingEvaluation resultLinguistic model
The invention discloses a real-time advertisement evaluation decision-making system based on a large language model agent. The system comprises a problem modeling module, a decision-making agent construction module and an advertisement evaluation result output module. The problem modeling module is used for modeling an original advertisement evaluation decision problem into a Markov process; the decision-making agent construction module constructs a decision-making agent based on a large language model and comprises a configuration file guiding unit, a click rate estimation model training unit, a basic decision-making factor generation unit, a memory module construction unit, an reflection module construction unit, an insight reasoning unit and an action making unit; and the advertisement evaluation result output module obtains a final advertisement evaluation result of each advertisement. According to the invention, the decision agent which integrates the click rate estimation model, the expert rule strategy, the multi-memory retrieval mechanism and the reflection mechanism is constructed based on the decision agent construction module to simulate the advertisement environment of the real world, so that the learning difficulty and the prediction difficulty of the model are reduced, and the efficiency and the effect of advertisement evaluation are improved.
Owner:SOUTH CHINA UNIV OF TECH
Driving test simulation control interaction method and related equipment
PendingCN119252107Aavoid position deviationImprove operational accuracyCosmonautic condition simulationsCharacter and pattern recognitionDriving testHand parts
The invention discloses a driving test simulation control interaction method and related equipment. The method comprises the steps that under the condition that driving test simulation is carried out in a closed VR visual environment, the position and posture of the hand of a user are captured in real time, and the user is located in a simulator cabin; mapping the acquired position and posture of the hand to a virtual environment of a VR visual environment, a virtual operation part in the virtual environment being consistent with a physical part for realizing the same function in the simulator cabin in position and proportion; and obtaining the operation of a user on the physical component and / or the virtual operation component so as to perform driving test simulation control in the simulator cabin. The problem that in driving test simulator training, although a VR simulator can really provide higher simulation reality and immersion, students cannot observe the real world when wearing VR glasses can be solved.
Owner:WUHAN FUTURE MIRAGE TECH CO LTD
Face forgery detection method based on identity decoupling and adaptive cosine embedding loss
This invention discloses a face forgery detection method based on identity decoupling and adaptive cosine embedding loss. Addressing the problems of existing deepfake detection methods being susceptible to interference from identity information and lacking flexibility due to fixed boundary constraints, this invention constructs an identity-independent deepfake detection network. It decouples the input face image into forensic features containing forgery clues and content features containing identity information, eliminating identity interference by ignoring content features. Simultaneously, it designs an adaptive cosine embedding loss based on visual differences, using absolute error and the visual differences between real and fake image pairs as adaptive boundary values to dynamically adjust feature spacing constraints. This invention significantly improves the model's generalization ability when facing unknown forgery types through the synergistic optimization of decoupled representation learning and adaptive metric learning.
Owner:ZHEJIANG GONGSHANG UNIVERSITY
Virtual object control method and related device
PendingCN119493556Aeasy to understandReduce learning difficultyVisual/graphical programmingImage data processingControl objectiveControl system
The embodiment of the invention discloses a virtual object control method and a related device, a simulation control system comprises a model building interface and a code editing interface, a virtual object is disassembled into a plurality of modules in advance, so that a user not only can independently build the virtual object in the model building interface, but also can build the code editing interface in advance; in addition, motion parameter configuration can be set for each virtual component included in the virtual object in the code editing interface to obtain a target virtual object. And when the user executes the running operation on the target virtual object, controlling the target virtual object to run based on the parameter configuration of the target virtual object. Therefore, a user not only can freely set up various module combinations to obtain the self-defined virtual object, but also can respectively control each virtual component included in the virtual object, so that the learning difficulty of the user in the programming process is reduced, and the creativity and the autonomous learning ability of the user are improved.
Owner:TENCENT TECHNOLOGY (SHENZHEN) CO LTD
Interface data processing method and system based on visual page and electronic equipment
PendingCN119668588AReduce learning difficultyLower operating thresholdError detection/correctionVisual/graphical programmingJSONComputer engineering
The invention relates to the technical field of interface management, and discloses an interface data processing method and system based on a visual page and electronic equipment. According to the application, a response parameter structure and a data processing rule of a target interface are respectively acquired through a visual page, a rule script is generated by adopting JSONPath according to the response parameter structure and the data processing rule which are converted into a JSON format, data processing is performed on a response parameter value of the target interface by utilizing the rule script, and a response result is obtained. Compared with development through a data processing tool and obtaining a response parameter structure and a data processing rule through a visual page, the operation threshold is reduced, meanwhile, data of different interface formats are unified through JSON and JSONPath, the data processing rule is unified for different interface types, and the data processing efficiency is improved. Therefore, the learning difficulty of interface data processing is reduced from two aspects, the learning cost is reduced, and the interface data processing efficiency is improved.
Owner:广域铭岛数字科技有限公司 +1
A language processing-based word vector representation method, device and terminal equipment
ActiveCN115983240BReduce the amount of parametersReduce learning difficultyNatural language data processingMachine learningAlgorithmTheoretical computer science
The application is suitable for the field of artificial intelligence technology, and provides a word vector representation method and device based on language processing and a terminal device, the method comprising: constructing an input word table; setting word frequency information of each word in the input word table; classifying the words into N clusters according to the word frequency information, words with the same word frequency information being classified into the same set to obtain a clustering set with N subsets; defining a first word vector dimension of the words in the first cluster through an experience factor, and obtaining a second word vector dimension of the words in each subset according to the subset where the words are located and the first word vector dimension; defining an nth linear mapping matrix for the nth cluster; and processing a target word vector through parameters shared by the input word table as an input of a natural language processing model. Through the application, the problem that the calculation amount is large, the efficiency is low and the word distribution features cannot be considered when converting words into word vectors in natural language processing can be solved.
Owner:SICHUAN LAN-BRIDGE INFORMATION TECHNOLOGY CO LTD
Training method, device and program product of language model
PendingCN122452765A
The specification provides a language model training method, device and program product. The training process of the language model is divided into multiple learning stages. The language model training method comprises: generating multiple constraint conditions for original question samples; based on the original question samples and the corresponding multiple constraint conditions, constructing multiple constraint question samples containing different numbers of constraint conditions; in each learning stage, generating a training set based on all the constructed constraint question samples; each learning stage corresponds to a target number, and the target number increases accordingly as the learning stage progresses; in the training set, the proportion of constraint question samples with the target number of constraint conditions is higher than that of other constraint question samples; training the current language model of each learning stage based on the training set of the learning stage; wherein, from the second learning stage, the current language model of any learning stage is the language model trained in the previous learning stage.
Owner:ANT BLOCKCHAIN TECHNOLOGY (SHANGHAI) CO LTD
Bus get-off station identification method and system based on CatBoost
The invention provides a CatBoost-based bus get-off station identification method and system, and the method comprises the steps: collecting and fusing bus station data, GPS data and IC card swiping data, and obtaining initial bus get-on and get-off station data through a trip chain algorithm; performing feature engineering processing on the bus IC card swiping data, including decomposing time column features, identifying workdays and holidays, and fusing geographic positions and external weather data; performing data preprocessing, and converting the data into a feature and label form for model training; building a CatBoost model based on a Python language, setting proper training parameters, and training and testing the model; and using the trained model to predict the data of the unrecognized getting-off stations. The method has the advantages that the recognition accuracy and credibility of the model are improved, and the model learning difficulty is reduced. The problems of low accuracy and low efficiency in traditional get-off station identification are effectively solved, and the method has good data processing capability and practicability.
Owner:XIHUA UNIV
Deployment method, inference method and electronic device of neural network model
This disclosure provides a method for deploying a neural network model, an inference method, and an electronic device. The deployment method includes: determining the chip platform corresponding to the neural network model; compiling a platform-level dynamic library and a device-level dynamic library corresponding to the chip platform, wherein the device-level dynamic library is configured to provide a first set of interfaces to the platform-level dynamic library, and the first set of interfaces is configured to call hardware-related interfaces of the chip platform; the platform-level dynamic library is configured to provide a second set of interfaces to the application layer, and the second set of interfaces is configured to call the first set of interfaces of the device-level dynamic library; and deploying the neural network model, the platform-level dynamic library, and the device-level dynamic library to the chip platform. This disclosure improves model deployment efficiency.
Owner:BEIJING SUNNIWELL DIGITAL S&T CO LTD
An automated testing system for flight controllers
ActiveCN224287402Ulow costReduce learning difficultyElectric testing/monitoringTotal factory controlControl cellInstrumentation
This utility model discloses an automatic testing system for flight controllers, relating to the field of unmanned aerial vehicle (UAV) technology. It includes a control unit, a sensor unit, a storage unit, a detection unit, a power supply unit, a display unit, an input unit, and an interface unit. The control power supply includes a master MCU and a slave MCU. The control unit is connected to the sensor unit, storage unit, detection unit, power supply unit, display unit, input unit, and interface unit. The sensor unit includes a compass, a barometer, and an IMU, and is connected to the detection unit. The storage unit contains a TF card. This utility model eliminates the need for additional instruments, reducing the cost of purchasing equipment. Furthermore, only the automatic testing system needs to be learned, reducing the learning difficulty and manual training costs for workers. Simultaneously, the flight controller is connected to the system via a single cable, reducing connection difficulty and shortening testing time from approximately 2 hours to about 10 minutes.
Owner:VIEWPRO LTD
Multifunctional acupuncture and cupping device
InactiveCN119606729Aimprove bindingReduce the difficulty of operationAcupuncturePneumatic massagePhysical therapySurgery
The invention relates to the technical field of acupuncture and cupping equipment, in particular to a multifunctional acupuncture and cupping device which comprises a cup body and further comprises a conductive assembly, an air exhaust assembly and a pressure relief assembly, and the conductive assembly is connected with the cup body and comprises a wire clamping assembly, a needle body clamping assembly and a middle connecting assembly. The needle body clamping assembly can be synchronously driven to be opened when the wire clamping assembly is opened, the air exhaust assembly is installed in the middle connecting assembly, the interior of the tank body can be vacuumized through the air exhaust assembly, meanwhile, the air exhaust assembly can be synchronously driven to operate when the wire clamping assembly is opened, and the pressure relief assembly is installed on the tank body. And vacuum in the tank body can be broken through the pressure relief assembly. According to the invention, a power supply line for electroacupuncture treatment can be directly and electrically connected with the needle body in the cup body through the conductive assembly, so that the electroacupuncture treatment does not need to be additionally performed around the cup body, and the operation difficulty and the learning difficulty are greatly reduced.
Owner:李洋
Energy consumption prediction method, device and system of air conditioning system
PendingCN121953428Aimprove accuracyImprove reliabilityMechanical apparatusNeural learning methodsPredictive methodsTerm memory
The invention provides an energy consumption prediction method, device and system for an air conditioning system. The accuracy and reliability of energy consumption prediction can be improved. The method comprises the steps of obtaining real-time operation state data of an air conditioning system of a building; a data driving model is obtained, the data driving model is obtained through training on the basis of historical operation state data and a physical mechanism model of the air conditioning system, and a neural network structure adopted by the data driving model is a long-short-term memory network; the physical mechanism model is constructed based on static physical parameters of the building, historical environmental data and a preset thermodynamic law; and the real-time operation state data is processed based on the data driving model, an energy consumption predicted value is obtained, and the energy consumption predicted value is used for generating a control instruction for adjusting the air conditioning system.
Owner:GUANGZHOU ELECTRONICS TECH +1
Section display method, device and equipment of three-dimensional figure and computer program product
PendingCN119418566AReduce learning difficultyData processing applicationsEducational modelsGraphicsAlgorithm
The invention provides a three-dimensional figure section display method, device and equipment and a computer program product, the method can determine a plane where input target points are located together, the number of the target points is at least three, intersection points generated by the plane where the target points are located together and all edges of a three-dimensional figure are calculated and connected, and the intersection points are obtained. The section graph is generated in the three-dimensional graph and is displayed, so that the section graph can be automatically generated in the three-dimensional graph according to the target point input by the user and is displayed, students can intuitively know the position of the section, and the learning difficulty of the students for solid geometry is reduced.
Owner:JIANGSU XUNZHI FUTURE INFORMATION TECHNOLOGY CO LTD
A robotic arm teleoperation interaction method based on digital twin of target object
The present invention relates to a robot arm teleoperation interaction method based on a digital twin of a target object, which generates a digital twin of a real object at a corresponding position of a virtual scene using a prefabricated three-dimensional model through object recognition and positioning, and can put a new prefabricated three-dimensional model into a model library for reconstructing the scene, while integrating the operation of the robot arm into the scene; the user operates the robot arm by directly interacting with the digital twin of the object to be grasped. The robot arm teleoperation interaction method based on a digital twin of a target object of the present invention is a direct interactive operation mode for the target object, which reduces unnecessary collisions caused by user operation errors, and reduces the workload and difficulty of use; the use of prefabricated three-dimensional models instead of point cloud models greatly reduces the amount of data that needs to be transmitted, ensures real-time performance, and improves user experience.
Owner:CHANGCHUN UNIV OF SCI & TECH
Handheld control device and multi-degree-of-freedom surgical instrument
PendingCN121943422AImprove efficiencyHigh precisionSurgical forcepsPhysical medicine and rehabilitationPhysical therapy
The invention relates to the field of surgical instruments, in particular to a handheld control device and a multi-degree-of-freedom surgical instrument. The deflection mechanism is rotationally connected to the holding body and provided with a first driving column and a second driving column, and the deflection mechanism can be shifted to drive the first driving column and the second driving column to deflect relative to the holding body; the transmission mechanism is connected to the first driving column and the second driving column so as to drive the first driving column and the second driving column to rotate. According to the device, the independent deflection mechanism and transmission mechanism are arranged, so that physical decoupling of two main actions of opening, closing and deflection of the tail end clamp body is realized; deflection and rotation are mutually independent, control paths are clearly separated through overall deflection or respective rotation of the driving column, the motion coupling phenomenon caused by linkage of driving wires in the prior art can be avoided, control is more accurate, the hand feeling is direct, extra compensation by a doctor is not needed, operation complexity is reduced, and operation efficiency is improved. And the continuity and the accuracy of the operation are ensured.
Owner:SHANGHAI FUYI MEDICAL TECHNOLOGY CO LTD
Wavelet-space dual attention image deraining method and system guided by prior knowledge
The present application relates to the technical field of image processing, and more particularly to a priori knowledge guided wavelet-space double attention image rain removal method and system, comprising: obtaining an original rainfall image as an initial input image; using a priori knowledge guided wavelet-space double attention rain removal deep learning network model to perform rain removal processing on the initial input image to obtain a rain-free image; the a priori knowledge guided wavelet-space double attention rain removal deep learning network model comprises an input mapping module, a priori knowledge guiding module, a wavelet-space double attention based hierarchical encoding-decoding network and an output mapping module. The rain removal deep learning network model used in the present application combines a priori knowledge guiding mechanism, wavelet-space double attention and multi-scale feature interaction strategy to enable the network model to better understand the structure and features of rainy day images, enhance feature extraction, improve image detail and texture quality, and reduce network learning difficulty.
Owner:SUN YAT SEN UNIV
Experimental device for exploring relationship between pressure of liquid on bottom of container and shape of container
ActiveCN222365323UReduce learning difficultyEasy to learnEducational modelsMechanical engineeringMechanics
Owner:HUNAN SHANJIAN TEACHING INSTRUMENT EQUIPMENT CO LTD
Image generation method based on gradient noise suppression generative adversarial network
ActiveCN119229192AImprove sampling efficiencyReduce computational complexityCharacter and pattern recognitionBiological modelsAlgorithmGenerative adversarial network
The invention discloses an image generation method based on a gradient noise suppression generative adversarial network, and the method comprises the following steps: obtaining a training set which is used for a positive sample of a discriminator; generating an image set as a negative sample; training a gradient noise sensor; training parameters are set in GNS-GAN adversarial learning, and weights of a discriminator and a generator are calculated and updated; and finally, generating an image by using a generator and carrying out pixel value scale exchange to finish the generation of a final image. The method has the advantages that the quality of the generated image and the stability of adversarial learning are improved, and the noise influence in the training process is effectively reduced through the gradient noise suppression technology; training time and computing resource consumption are reduced, and complex forward and reverse diffusion steps in a traditional diffusion model are avoided; the network structure is improved, the parameter quantity of the generator and the discriminator is less, and the training efficiency of the model is improved.
Owner:SOUTHWEAT UNIV OF SCI & TECH
Power Internet of Things terminal access parameter adaptive optimization method and system based on hierarchical reinforcement learning
PendingCN121967211Aefficient executionWill not fall into short-sightednessTransmissionAdaptive optimizationTerminal equipment
The invention discloses a power Internet of Things terminal access parameter adaptive optimization method and system based on hierarchical reinforcement learning, and the method comprises the steps: carrying out the optimization modeling of a terminal access parameter into a hierarchical reinforcement learning decision process, and carrying out the access strategy selection and parameter fine-grained adjustment through a high-layer intelligent agent and a low-layer intelligent agent, and in combination with Monte Carlo sampling, simulation evaluation is carried out on a dynamic network environment, so that online adaptive optimization of access parameters of the terminal equipment is realized. In consideration of diversity and complexity faced by equipment access in an electric power internet of things system, the access parameters of the equipment are optimized in real time by dynamically analyzing the running state and service requirements of the terminal equipment and combining five modules including a sampler, a parameter judgment device, an object model state machine, a benchmark tester and a reward model. By adopting the method provided by the invention, challenges such as terminal large-scale access and heterogeneous network environment dynamic change in the power internet of things can be effectively handled, and the terminal access reliability, the network throughput and the communication efficiency are remarkably improved.
Owner:GUANGDONG POWER GRID CO LTD +1
A tunnel clearance convergence measurement method
The tunnel clearance convergence measurement method described in the present invention is applied to the measurement and calculation of the convergence displacement of the tunnel structure when the tunnel support structure or the surrounding rock and soil body is offset, so as to effectively solve the purpose of accurate measurement and calculation after the rock and soil body is offset, and realize the high-precision requirement of tunnel clearance convergence measurement. The tunnel clearance convergence measurement method sets up a total station at an observation point, and uses the total station to measure the distance a1 from the total station to the left clearance convergence point and the distance a2 from the total station to the right clearance convergence point, the angle α1 of the rear sight point-total station eyepiece-left clearance convergence point, and the angle α2 of the rear sight point-total station eyepiece-right clearance convergence point; uses a laser rangefinder to measure the height c of the total station, the distance b1 from the observation point to the left clearance convergence point, and the distance b2 from the observation point to the left clearance convergence point; obtains the clearance convergence value l after vertical and horizontal correction through observation, and the clearance convergence value l after correction minus the clearance convergence initial value l0 is the cumulative change of tunnel clearance convergence.
Owner:WUHAN YANGTZE RIVER TUNNEL CONSTRUCTION CO LTD
A proximal humerus fracture Neer classification teaching system and method
The application relates to the technical field of medical teaching, in particular to a proximal humerus fracture Neer classification teaching system and method, wherein the system comprises: a data mapping module, which is used for automatically identifying a finger state and analyzing the finger state into a fracture displacement parameter of corresponding proximal humerus anatomical structure according to a mapping relationship; and a rendering simulation module, which is used for driving a three-dimensional proximal humerus fracture model according to the fracture displacement parameter, superimposing the three-dimensional proximal humerus fracture model on the fingers or presenting the three-dimensional proximal humerus fracture model in a virtual space for rendering simulation in a visual environment. The application simulates fracture displacement according to a mapping relationship and a gesture rule by establishing a mapping relationship between the fingers and the anatomical structure, converts abstract medical classification into physical actions that can be personally perceived and operated by learners, and thus greatly reduces learning difficulty and improves understanding depth and memory persistence.
Owner:NANFANG HOSPITAL OF SOUTHERN MEDICAL UNIV
Method and device for identifying interaction among components of compound
PendingCN119649953AReduce learning difficultyImprove learning efficiencyChemical machine learningChemometricsUser inputEngineering
The embodiment of the invention relates to a method and a device for identifying interaction among compound components. The method comprises the following steps: receiving a compound structure file input by a user as a corresponding first compound and storing the first compound; receiving an interaction screening type input by a user and two corresponding different types of component identifiers as a corresponding first screening type and a first group of component identifiers; and identifying whether the interaction relationship between the two compound components corresponding to the first group of element identifiers and the first screening type is established or not according to the first compound to obtain a corresponding interaction relationship identification result, and feeding back the result to the current user. A friendly basic recognition tool can be provided for workers in the chemical field, the workers are helped to reduce the learning difficulty, the learning efficiency is improved, and effective assistance is provided for the workers to master knowledge in the field as soon as possible.
Owner:BEIJING DP TECH CO LTD
A universal verification system for integrated circuit chip verification
ActiveCN119201097BImprove verification efficiencyImprove collaborationError detection/correctionSoftware reuseProgramming languageVerification
The present invention belongs to the field of digital chips, and relates to a universal verification system for integrated circuit chip verification, wherein the top-level code directory includes a bin directory, a lib directory, and a pub directory, and the directory of each verification platform of the universal verification system includes a tb directory, an env directory, a testcases directory, an etc directory, and an inst.fl directory; after the simulation starts, all configuration information under each etc directory is collected, and then the compilation parameters inherited by the corresponding test case are selected according to the verification case to be simulated specified by the user to compile the verification platform; after the compilation is completed, the simulation parameters specified by the user are added for simulation; after the simulation is completed, the log file of the simulation is checked according to the inspection rules specified in the build to determine whether the verification case is passed. The present invention has the advantages of unified verification environment, flexible test case management, real-time verification result display, low learning cost, high resource utilization, cross-module collaboration, unified parameter management, and strong versatility.
Owner:BEIJING SHUDU INFORMATION TECH CO LTD