115results about How to "Avoid consistency" patented technology

The invention relates to the technical field of novel power system optimization operation, in particular to a new energy high-uncertainty scene generation method and device based on a CWGAN-GP model, and the method comprises the steps: correcting an initial generative adversarial network through employing a Wasserstein distance and a gradient penalty term, so as to obtain a final generative adversarial network, and the target convolutional neural network is introduced into the final generative adversarial network to obtain a conditional deep convolutional generative adversarial network of improved deep convolution, so that WGAN-GP is introduced into the conditional deep convolutional generative adversarial network of improved deep convolution to obtain a CWGAN-GP model of improved deep convolution. And a target L-ISODATA clustering algorithm is applied to cluster the generated wind and light scenes so as to generate a high-uncertainty new energy power generation output scene. Therefore, the problems that in the related technology, when the GAN processes high-dimensional data, complex space-time characteristics are difficult to effectively capture, and due to lack of prior information, the training direction is prone to deviation, and the scene generation effect and stability are affected are solved.

The invention discloses a visual detection optimization method and system based on data set classification. The method comprises the steps of introducing a scene classification label and carrying out standardization processing on scene features in a data set labeling and preprocessing stage, and carrying out scene classification multi-task learning in a model selection and training stage, so that a model dynamically adjusts feature learning in different scenes, and the multi-scene adaptive capacity is improved. And in the model evaluation and verification stage, scene-based hierarchical evaluation is introduced and corresponding optimization is carried out. In the model deployment and reasoning stage, a dynamic scene is introduced to adapt to reasoning, and the reasoning path of the model is dynamically adjusted to adapt to the characteristics of the scene, so that the reasoning accuracy is improved. In the later optimization, scene feature clustering and transfer learning are used, and the multi-scene performance of the model is optimized. According to the method, under the condition that the model is not jointly trained through multiple data sets, the multi-scene adaptive capacity of the visual detection model is improved through data set classification.

The invention relates to the technical field of emotion recognition, in particular to a multi-modal emotion recognition method based on deep learning, which comprises the following steps: constructing and training a multi-modal emotion recognition model, and inputting to-be-recognized data into the trained multi-modal emotion recognition model to output a recognition result; the multi-modal emotion recognition model comprises a self-attention encoder, a first cross attention encoder, a second cross attention encoder, a gating unit and an MLP classifier; according to the method, comprehensive learning and interaction are carried out by combining two modal data of voice and text, the emotion information is more comprehensive, the inherent defect of a single feature is made up, and the robustness of the model is higher and the accuracy is higher.

The invention relates to the field of vehicle network communication, in particular to an automobile attack chain analysis method and device, electronic equipment and a storage medium. According to the method, the security event and the corresponding event data are firstly obtained to provide a basis for subsequent analysis; then, safety key elements in the event data are extracted, a knowledge graph is constructed with the safety key elements and the automobile software and hardware assets, the safety key elements and the automobile software and hardware assets serve as different entity nodes respectively, the association between the different elements is determined in this way, and the problem that the association of attack behaviors is neglected is solved. Then, analyzing the knowledge graph to obtain target malicious nodes to construct an initial attack graph, and analyzing event data to obtain an attack link to update a node relationship of the initial attack graph to obtain a target attack graph, so that a unified description framework is provided for complex attacks; the whole attack process and the relation between the nodes can be clearly known, and the inconsistency of understanding and reaction caused by lack of a unified description framework is avoided.

The invention discloses a virtual reality fusion method based on an unmanned aerial vehicle video, and relates to the technical field of virtual reality, and the method comprises the steps: predicting the current and next concern points of a user through eye movement tracking and Kalman filtering, preferentially distributing rendering resources to the areas, and employing a dynamic resource distribution mechanism to achieve the fusion of the current and next concern points of the user; according to the method, resources are allocated to a region about to be concerned by a user in advance, meanwhile, the processing priority of a secondary region is reduced, waste of computing resources is avoided, meanwhile, the moving track of a fixation point of the user is predicted through a Kalman filter, the next concern point of the user is rendered in advance, rendering of the concern region is completed when the visual angle is rapidly switched, and the user experience is improved. Compared with the prior art, the method has the advantages that image lagging and blurring felt by a user are reduced, rendering efficiency is greatly improved, user experience is smoother and more natural, a multi-level cache framework is adopted, the problem of loading delay in a traditional method is avoided, and the response speed of a system is effectively increased.

The invention discloses a gel electrolyte raw material high-temperature mixing device, and relates to the technical field of stirring and mixing, the gel electrolyte raw material high-temperature mixing device comprises a base, the base is fixedly connected with a stirring barrel, the stirring barrel is fixedly connected with a cover plate, the inner bottom of the stirring barrel is fixedly connected with a ring-shaped heating plate, the stirring barrel is internally fixedly connected with a conversion cylinder, and the conversion cylinder is fixedly connected with the cover plate. Through holes which are evenly distributed in the circumferential direction are formed in the side, close to the heating plate, of the conversion cylinder. On the basis of stirring and mixing of the first stirring blades, raw materials flow up and down and are mixed through the spiral conveying frame, the uniform mixing degree of the raw materials is enhanced, the raw materials flow to the low-temperature area through the high-temperature area, the raw materials are integrally and synchronously heated, and local overheating caused by uneven heating is avoided; therefore, thermal decomposition or chemical reaction initiation of some raw materials is avoided, harmful by-products are prevented from being generated, the chemical stability and electrochemical performance of the electrolyte are prevented from being influenced, meanwhile, the phenomena of phase separation and inconsistent viscosity change caused by temperature difference are avoided, and the formation of a non-uniform mixture is avoided.

The invention relates to the technical field of artificial intelligence, in particular to an intelligent large-screen multi-round dialogue linkage method based on a large model. A user can control the intelligent large screen anytime and anywhere through the mobile phone without being limited by time and equipment operation. Compared with a traditional interaction mode, the user experience is more natural and convenient. A two-dimensional code generated by an intelligent large screen is scanned through a mobile phone program, a safe interaction channel is established through a WSS protocol of Websocket, and multi-round dialogue control is performed by using a large model.

The invention discloses a mahjong matching game level generation method based on reinforcement learning, a medium and equipment, and belongs to the technical field of machine learning, game levels are randomly generated, then game problem solving is executed based on a reinforcement learning network model, the level passing rate is obtained, and if the level passing rate is larger than or equal to a set threshold value, the game levels are output. In the reinforcement learning network model, a neural network is used as a strategy network and a value network to extract level features, and the value network cooperates and simulates the time required for playing of a real player and the level passing rate. The game level can be generated according to the preset difficulty and layout of the game, so that the game has the level design of diversified card surface types, quantity and arrangement modes under each difficulty level, and the playability and challenge of the game are remarkably improved. According to the invention, level challenge experience with invariable difficulty can be provided, and the problem of inconsistent user experience caused by inconsistent difficulty due to random generation of levels is avoided.

The utility model discloses a bending device for bracket processing, which relates to the technical field of machining, and comprises a workbench and a bending mechanism, a bracket needing to be bent is clamped and fixed through a fixing frame, in the bracket bending process, a rotating shaft is driven by a motor to further drive the bracket to move and feed, and the bracket is bent by adjusting the length of a telescopic pipe. In the feeding process, impurities on the surface of the support are scraped off through a friction scraping plate, a support cleaning agent stored in a liquid storage box enters the limiting blocks through telescopic pipes and is sprayed to the surface of the support through through holes, stubborn impurities on the surface of the support are further removed, and the cleaning effect of the support is improved. The problem that the support is bent unevenly or deformed due to impurities is solved, the support is heated through the electromagnetic heating plate, the hardness and strength of materials are reduced, meanwhile, a cleaning agent is dried, and it is prevented that the cleaning agent stays on the surface of the support, secondary pollution is caused to the support, and the quality of the support is affected.

The invention relates to the technical field of stainless steel thin strip machining, in particular to stainless steel thin strip cutting equipment which comprises a machine body, an unwinding mechanism, a cutting mechanism, a grinding mechanism and a collecting mechanism are arranged on the machine body, the unwinding mechanism is located at the front end of the machine body and comprises an unwinding motor and an unwinding roller, and the unwinding roller is rotationally borne on the upper surface of the machine body; the unwinding motor drives the unwinding roller to rotate, the cutting mechanism comprises a first crank-link mechanism and a cutter, a rotary disc of the first crank-link mechanism is rotationally borne above the machine body, the cutter is arranged above the machine body in a sliding mode in the vertical direction, a connecting rod of the first crank-link mechanism is hinged to the cutter, and the grinding mechanism is located at the rear end of the cutting mechanism. The grinding mechanism is used for grinding the cut edge of the stainless steel thin strip, and the collecting mechanism is located at the tail end of the machine body and used for collecting the stainless steel thin strip; the device has the advantage that the production efficiency, precision and safety of the stainless steel thin strip are improved.

The invention discloses a specimen sampling device based on medical detection, and relates to the technical field of blood sampling, the specimen sampling device comprises a vacuum blood collection tube, a blood collection needle and a fixator, the vacuum blood collection tube comprises a bottle tube and a bottle cap, the fixator comprises a positioning frame, a sliding block, a puncture needle, a first spring, a second spring and a buckle, and a mounting groove is formed in the upper end face of the positioning frame; the vacuum blood collection tube is detachably installed in the installation groove, the sliding block is installed in the installation groove and is in elastic sliding connection with the positioning frame through a first spring, the puncture needle is installed in the sliding block and is in elastic sliding connection with the positioning frame through a second spring, and a hose is arranged between the blood collection needle and the puncture needle. The blood sampling device has the advantages of being convenient to operate and capable of automatically mixing the anticoagulant, and the problems that according to existing blood sampling equipment, medical staff need to operate with two hands, the blood sampling part of a patient is fixed with one hand, operation is complex, misoperation is prone to occurring, and the process of manually replacing a blood sampling bottle and shaking and mixing the anticoagulant is long in time consumption and low in efficiency are solved.

The invention relates to the technical field of computers, and discloses a file verification method and device for a content distribution network, equipment, a medium and a product, and the method comprises the steps that a source station carries out abstract operation on a target file to obtain a first abstract file, and constructs a first protection file of the first abstract file; wherein the filename of the first protection file is obtained by encrypting the filename of the first abstract file based on a first encryption key corresponding to the target file, and the first protection file is used for verifying the accuracy of the first abstract file. In response to an access request for a target file, the source station sends the target file, the first digest file, and the first protection file to the content distribution network. And the content distribution network performs accuracy verification on the received target file based on the first abstract file and the first protection file. According to the invention, the problem of file transmission errors caused by tampering of the file transmitted by the content distribution network can be solved.

The utility model belongs to the technical field of VOCs monitoring and detection. The utility model provides a particle state pretreatment device. The particle state pretreatment device comprises an upper part, a lower part, an annular sealing ring, a quartz filter membrane and a stainless steel filter membrane net support, the upper part comprises a first port and a second port; the first port is formed in the center of the top end of the upper part; the second port is formed in the side wall of the upper part and is close to the top end; the lower part comprises a third port and a groove buckle; the third port is formed in the center of the bottom end of the lower part; the groove buckle is arranged on the inner side of the top end of the lower part; the stainless steel filter membrane net support is placed in the groove, the quartz filter membrane is placed on the stainless steel filter membrane net support, and the annular sealing gasket is placed on the quartz filter membrane and used for sealing the connecting position of the upper portion and the lower portion. According to the device provided by the utility model, on the basis of separating a dissolved state from a particle state, the actual concentration of the VOCs in water can be really reflected by monitoring data, and the problem of lack of the particle VOCs in an existing online method is avoided.

The invention discloses an RTK high-precision positioning tamper construction data management method and system, and relates to the field of tamper construction in composite foundation engineering, and the method comprises the following steps: obtaining the GPS positioning data of different working parts of a tamper through a single-machine double-GPS module, and positioning the respective position and height; processing the obtained GPS positioning data by using an RTK technology to obtain high-precision position information; and according to the high-precision position information, whether the position of the rammer deviates from the ramming pile position or not is judged in combination with construction pile position distribution diagram data, and if deviation occurs, alarm information is sent out. Through the design of the single-machine double-GPS module, the position and height information of the rammer body and the rammer can be obtained at the same time, data synchronism is guaranteed, the problems that cost is increased and information synchronization is inconsistent due to the fact that a plurality of devices are used are solved, the overall performance and stability of the devices are improved, and the practicability is high. And the requirement on multi-position information in the construction process is better met.

The invention relates to the technical field of building informatization, and discloses a building information model automatic generation and collaborative editing system which comprises a front-end user interface, a rear-end server, a core engine and middleware. The front-end user interface comprises a graphical user interface and an authority management module; the back-end server comprises a data storage module, an API (Application Program Interface) and a message queue module; the core engine comprises an automatic generation module and a collaborative editing module; the middleware comprises a file processing module and a cloud service integration module. Compared with the prior art, the method has the advantages that through the functions of automatic generation, collaborative editing, intelligent optimization, high compatibility, usability, visual display and the like, the building and editing efficiency of the BIM model is remarkably improved, the labor cost is reduced, meanwhile, collaborative work among multiple professional teams is achieved, and the efficiency of building and editing the BIM model is improved. And the project management efficiency and quality are improved. The method has wide application prospects and market competitiveness in the field of building informatization.

The embodiment of the invention provides a multimedia editing method and device and electronic equipment, and the method comprises the steps: generating initial data layer data according to the editing operation of a multimedia material, and the initial data layer data comprises multimedia material track fragment data obtained through the editing operation; determining an editing effect attribute indicated by the editing operation, associating the editing effect attribute with the initial data layer data, and taking the initial data layer data associated with the editing effect attribute as target data layer data; and according to a received display instruction, acquiring multimedia material track fragment data and the editing effect attribute from the target data layer data by utilizing a consumption view layer, and generating multimedia fragment content after editing operation based on the multimedia material track fragment data and the editing effect attribute. And data redundancy occurring in the multimedia editing process is reduced.

The invention provides a dual-frequency single-port intermodulation tester, and the tester comprises an intermodulation control module which comprises an intermodulation controller, a first frequency source and a second frequency source, and the intermodulation controller controls the first frequency source and the second frequency source to generate two carrier signals with different frequencies through receiving a configuration instruction; and the multiplexer module comprises a plurality of power amplifier modules and a multiplexer, the plurality of power amplifier modules are used for amplifying the two paths of carrier signals, synthesizing the two paths of carrier signals through the multiplexer, outputting the synthesized two paths of carrier signals to a tested piece, and receiving and processing intermodulation products reflected back from the tested piece at the same time. According to the invention, the accuracy, the integration level and the field applicability of the intermodulation test are effectively improved.

The invention discloses a delivery control system and method based on a large model and a sound cloning technology, and relates to the technical field of text delivery. Texts needing to be delivered in history are classified, a template of each text is generated, and a template database is constructed; inputting user data into the text template to automatically generate a delivery text, and constructing a text generation large model; selecting the receiving channel with the highest comprehensive efficiency as a conveying channel, processing the text according to the type of the conveying channel, and conveying the text to a delivery target; generating an evidence chain by using the data in the delivery process, and judging whether the delivery task is successful or not according to the generated evidence chain; carrying out secondary conveying by utilizing the optimized conveying channel, and carrying out secondary judgment on a conveying task; and communicating with the delivery target by using the personalized voice, explaining the text content to the delivery target, and performing opinion feedback after the delivery target clarifies the text content.

The application discloses a kind of based on edge computing's safe SIS instrument system interlocking control method, comprising the following steps: collecting and pre-processing the data in corresponding industrial field of safe SIS instrument system;Spatial correlation is constructed, and consistent constraint check, dynamic confidence evaluation and fusion processing are executed;Edge collaborative control architecture is constructed;Dismantle preset safety interlocking rule, and rule unit interlocking configuration result is configured;Distributable rule unit is distributed, and distributed interlocking strategy set is generated;Local interlocking decision result is generated and broadcast to the rest edge node;Global consistent interlocking decision result is generated and interlocking control instruction is output to corresponding executing agency.The application adopts edge collaborative interlocking control method, realizes multi-source data fusion and multi-node decision synchronization, with the advantages of high consistency, high reliability and fast response.

A transport system, a positioning device, a positioning method, a computer program product and a computer-readable recording medium. The positioning device comprises: a plurality of electronic tags arranged to be arranged at equal intervals along a line segment; and a fixed-frequency sampling module arranged to be able to move around the line segment. The fixed-frequency sampling module has two antenna parts and a positioning part. The positioning part is electrically connected to the two antenna parts. The two antenna parts couple the plurality of electronic tags and read the wireless radio frequency identification signals from the plurality of electronic tags at a fixed frequency. The positioning part estimates a coordinate on the line segment according to a plurality of identifiers and a plurality of intensity distribution curves from the wireless radio frequency identification signals. The antenna spacing between the two antenna parts is: d2=(n+0.5)×d1, d1 is a tag spacing between two adjacent electronic tags, n is an integer, and n≧0.

This invention provides a model context protocol invocation method and related apparatus based on edge-cloud collaboration, relating to the field of distributed system technology in cloud computing and edge computing. It generates a rollbackable workflow through a cloud-based MCP-Client, establishes a WebSocket connection with the user terminal, and sequentially sends multiple execution steps in the rollbackable workflow to the user terminal one step at a time. For each execution step, if the execution result is a failure, it rolls back to a specified rollback anchor point and dynamically replans the unexecuted steps based on the current state. This invention allows the cloud to dynamically adjust subsequent execution plans based on intermediate results through independently executable steps. The rollbackable workflow data format reduces the overhead of manual intervention and full retries, avoiding execution deviations and data inconsistencies. The use of WebSocket reduces latency and improves the real-time performance and interactivity of task scheduling.

The application discloses a kind of digital twin modeling and solving method of five-axis galvanometer laser processing system, belong to laser precision machining technical field, this method is based on the installation geometric parameter of four galvanometers, establishes the positive kinematics model of galvanometer optical path from galvanometer deflection angle to field mirror incident side light ray four-dimensional state;Again with light ray tracing simulation sample training by field mirror emission side four-dimensional state predicts incident side four-dimensional state reverse field mirror proxy model;When processing, focal plane target is converted into four-dimensional state containing position coordinate and direction cosine, after geometric tracing, proxy model prediction, with positive kinematics model as forward function to carry out numerical inverse solution, obtain four galvanometer deflection angle command.This method does not need to rely on f-theta field mirror internal prescription parameter, can realize the high-precision collaborative solution of focal point position and incident direction, while supporting closed loop check, sensitivity analysis, installation error disturbance simulation and four-dimensional residual compensation, applicable to variable incident angle laser precision machining scene.

The embodiment of the invention provides a social media platform interaction data display method and device based on Tab switching and electronic equipment, and belongs to the technical field of data processing.The method comprises the steps that user interaction data are obtained through an API construction tool, and standard data are obtained through preprocessing and stored; obtaining a feature vector set according to m feature value calculation functions, sorting the feature vector set according to time, and performing time sequence feature calculation to obtain a new feature vector; the standard data are clustered, and part of feature values are screened out through matching degree calculation; and constructing p different style display components for the screened characteristic values, and displaying the standard data in p paging pages through Tab switching to realize multi-dimensional display and interaction of the data. According to the method, an efficient data processing and transmission mechanism is adopted, real-time updating and displaying of the data are achieved, timely, accurate and comprehensive data analysis results are provided for users, and the social media platform data visualization effect and the user experience are remarkably improved.

The present disclosure relates to the field of computer technology, and discloses a file verification method, device, equipment, medium and product for a content distribution network. The method includes: a source station performs a digest operation on a target file to obtain a first digest file, and constructs a first protection file for the first digest file. The file name of the first protection file is obtained by encrypting the file name of the first digest file based on a first encryption key corresponding to the target file, and the first protection file is used to verify the accuracy of the first digest file. The source station sends the target file, the first digest file and the first protection file to the content distribution network in response to an access request for the target file. The content distribution network verifies the accuracy of the received target file based on the first digest file and the first protection file. The present disclosure can solve the problem of file transmission errors caused by tampering with files transmitted by a content distribution network.

According to the automatic spraying equipment for the hub low-pressure casting mold, a diaphragm pump is connected with a coating supply pipeline I and extracts coating from a coating barrel, the extracted coating can flow back into the coating barrel again through a color changing device to form a supply loop, and the color changing device can automatically select the needed coating type through an electric control system; the mold temperature measuring device is used for measuring the temperature of the low-pressure casting mold placed on the mold positioning platform in real time and feeding back temperature data to the electric control system, the laser red line emitter is used for emitting a positioning red line, and when the mold positioning platform returns to the original point, the laser red line emitter emits the positioning red line. The center of the positioning red line coincides with the center of the mold positioning platform. In the automatic spraying process of the hub low-pressure casting mold, the problems of closed-loop supply of the coating, full-size mold positioning and automatic temperature monitoring in the mold spraying process can be solved, and manual mold coating spraying work can be replaced.

The invention relates to the related technical field of microbial fermentation, and discloses microbial fermentation culture equipment for enhancing the sterile culture effect, which comprises a fermentation tank, the lower end of the fermentation tank is provided with a tank bottom, the upper end of the fermentation tank is provided with a tank cover, the fermentation tank is internally provided with a stirring assembly, the tank cover is internally provided with an air inlet assembly, and one side of the tank cover is provided with an inlet. A round pipe is rotationally arranged in the fermentation tank; according to the device, the flowing path of gas on the upper side in the shell can be prolonged through the wave-shaped channel, so that the gas is in full contact with an adsorption material, and the gas impurity removal effect is improved; then the fan-shaped sliding plate moves downwards, so that the adsorption material in the wave-shaped channel falls to the lower side of the interior of the shell to be mixed; and finally, through vertical shaking of the fan-shaped sliding plate, the shaking and mixing effect of the adsorption materials is improved, the adsorption materials in the wave-shaped channel can be redistributed, the adsorption materials in the wave-shaped channel can make full contact with gas, and the service life of the adsorption materials is prolonged.

This utility model discloses a miniature ball bearing for a cooling fan, comprising an outer ring, an inner ring, a cage, a dust cover, and rolling elements movably disposed between the outer and inner rings. Both the inner and outer peripheral walls of the outer ring and the inner ring have circumferentially formed grooves for the rolling elements to move. Both ends of the inner peripheral wall of the outer ring have circumferentially formed stepped grooves for installing the dust cover. The connections between the grooves of the inner ring and the outer peripheral wall of the inner ring, and between the grooves of the outer ring and the inner peripheral wall of the outer ring, have chamfered edges. A retaining spring is provided in the stepped groove for an interference fit with the stepped groove. The inner wall of the retaining spring abuts against the outer wall of the dust cover, while the outer peripheral wall of the dust cover has a clearance fit with the inner peripheral wall of the stepped groove. The curvature coefficient of the grooves of the outer ring is greater than that of the grooves of the inner ring. This utility model solves the problems of traditional miniature ball bearings, such as increased bearing ellipticity due to dust cover installation leading to out-of-tolerance bearing ellipticity, and assembly damage during rolling element installation leading to unqualified rolling element vibration during bearing operation.

The present invention discloses an application analysis and supervision system and method for compatible software architectures, and relates to the technical field of compatible software architecture applications. The present invention configures and deploys a thread supervision API and creates an execution lock to monitor the execution progress of tasks in different software architectures in real time, ensure the synchronous execution of tasks in a multi-architecture environment, avoid resource competition and scheduling failure, judge and manage the migration of threads between different software architectures, ensure the efficiency and accuracy of task execution, reduce the complexity of system integration, ensure the synchronization of task execution and data consistency, perform thread migration judgment and status tracking in a multi-architecture environment, effectively improve the execution accuracy of the system, avoid resource waste caused by insufficient resources or large differences in software architecture performance, ensure the coordination of tasks in different software architectures, support the expansion of different software architectures, improve the flexibility and adaptability of the system, and adapt to diversified computing platform requirements.

The application discloses an emotion speech generation method, system, device, storage medium and program product, relates to the technical field of speech synthesis, and comprises the following steps: obtaining target emotion parameters, and determining corresponding target emotion coordinates in a three-dimensional continuous emotion dimension space; generating candidate speech by using a speech synthesis model; determining an emotion reward value based on the deviation between the emotion coordinates of the candidate speech in the three-dimensional continuous emotion dimension space and the target emotion coordinates, in combination with a tolerance margin and a weight vector of each emotion dimension; and performing strategy optimization on the speech synthesis model based on the emotion reward value and a content reward value and / or a sound quality reward value, so as to generate speech corresponding to an emotion intensity. The scheme is based on a continuous VAD emotion space, realizes decoupled fine regulation and control of the emotion intensity, generates high-fidelity and highly controllable emotion speech through an emotion reward guide model, takes into account naturalness and content intelligibility, and avoids distortion and inconsistent expression caused by discrete labels.