33results about How to "Improve real-time response performance" patented technology

The invention belongs to the technical field of underwater acoustic communication reconnaissance and artificial intelligence edge computing, and discloses a sea area signal reconnaissance method and device based on deep learning. According to the method, a constructed lightweight LeNet-5 hierarchical convolutional neural network model is embedded and deployed into an STM32H743 microcontroller through an X-Cube-AI tool to process underwater acoustic signals, automatic identification and classified modulation are performed on underwater signals of active sonar signals and non-cooperative communication signals in real time, and the underwater signals are used for analysis of recovered sea area information. According to the method, deep-level features can be automatically learned from the time-frequency graph, better robustness is achieved for noise and a complex marine environment, and the recognition accuracy is higher.

This invention relates to the field of equipment training technology, and discloses an electro-dynamic feedback device and design method for virtual training, including the following steps: Designing an arm-worn force feedback device based on a DC servo motor; establishing a mapping relationship between virtual environment interaction information and physical feedback force by deriving a mathematical model of feedback force and servo motor rotation angle; designing a phase displacement-torque dual closed-loop control system, controlling the servo motor rotation angle through a phase displacement closed loop and controlling the servo motor output torque through a torque closed loop to achieve precise control of the feedback force amplitude; using the mathematical model and the dual closed-loop control system, driving the arm-worn force feedback device to generate feedback force matching the virtual training scene, achieving rapid response to virtual collisions, and keeping the feedback force and rotation angle error within a preset threshold to enhance the realism and immersion of virtual training. Establishing a mapping relationship between virtual interaction and physical feedback enables precise control of the servo motor rotation angle and output torque.

The application belongs to the technical field of artificial intelligence driven system design optimization, and discloses a training course planning service method and system. A quantum principal component analysis module is used to extract a correlation feature matrix of scene fluctuation parameters and resource demand, a quantum Fourier transform is used to analyze system health degree attenuation signals, and dynamic scene features are generated. A parameter tuning instruction set is constructed based on a quantum annealing algorithm to drive a genetic algorithm to optimize the topology. A quantum superposition state mapping is used to distribute micro-optimization units to a digital twin platform, and a variation quantum algorithm is used by a collaborative verification module to solve resource multiplexing strategies to form a data closed-loop feedback link. The application solves the problem of static data and dynamic demand mismatch, and improves real-time responsiveness and resource optimization efficiency in complex scenarios.

The application provides a multi-layer satellite network distributed routing method and system, the method comprising: dividing a low-orbit satellite set into multiple low-orbit satellite regions, each low-orbit satellite region corresponding to at least one medium-high-orbit satellite as a regional control node; multiple medium-high-orbit satellites cooperatively generating an exit selection strategy for each low-orbit satellite region and issuing the exit selection strategy to the corresponding low-orbit satellite region; low-orbit satellites in a low-orbit satellite region taking the corresponding exit selection strategy as an inter-regional forwarding constraint and performing hop-by-hop autonomous forwarding in the low-orbit satellite region; and multiple medium-high-orbit satellites updating the exit selection strategy for each low-orbit satellite region based on updated regional network state. The multi-layer satellite network distributed routing method and system of the application can reduce the frequency of cross-region path reconfiguration while retaining the real-time response capability of hop-by-hop routing, thereby enhancing the overall stability and scalability of the system.

The application provides a communication method, an electronic device, a storage medium, a power supply, a lamp and a lighting system. The communication method is applied to a host computer, the host computer is connected with a plurality of slave computers, and the method comprises the following steps: a broadcast packet is sent to all the slave computers, the broadcast packet is used to indicate that the host computer is ready to extract information; after the broadcast packet is sent, a plurality of first cycles are performed; the plurality of first cycles are ended, and a plurality of information packets are accepted, the information packets are from the slave computers which have a message sending intention; wherein the first cycle comprises the following steps: a plurality of reporting identifiers are received, the plurality of reporting identifiers are from a subset of all the slave computers which have a message sending intention; an identification is formed according to the plurality of reporting identifiers, and the identification is sent to the slave computers in the subset. When the number of slave computer devices is large, the real-time response capability of the host computer is improved, and the experience of the master-slave system is optimized.

The invention discloses a multi-modal heterogeneous data increment embedding fusion method, equipment and a medium, mainly relates to the technical field of multi-modal heterogeneous data, and aims to solve the problems of difficult neighborhood structure maintenance, insufficient cross-modal information fusion, single and fixed index structure and the like in the existing scheme. Comprising the following steps: converting modal data in a multi-modal heterogeneous data set into hidden space representation, and fusing the hidden space representation by utilizing a cross-modal attention mechanism to obtain historical embedding representation of a preset d dimension; when the newly-added multi-mode heterogeneous data set appears, calculating a newly-added embedding representation of a preset d dimension; according to the newly-added multi-modal heterogeneous data set, the newly-added embedding representation, the historical multi-modal heterogeneous data set and the historical embedding representation, constructing an adjacency relation graph of the newly-added multi-modal heterogeneous data set; and inputting the adjacency relation graph into a pre-trained incremental graph neural network to obtain an embedded representation fusing the adjacency relation.

The application discloses a kind of multi-modal event primitive extraction, coding and collaborative generation method and system for smart engineering construction scene, the system includes multi-modal data acquisition module, edge preprocessing module, event candidate extraction module, time-space alignment module, subject association module, event primitive coding module, multi-source fusion generation module and primitive output module. Through edge pre-processing, candidate event extraction, time alignment, space mapping, subject association, unified coding and multi-source fusion to video, environmental parameters, device status, positioning information, wearable terminal data and job log data and other multi-modal raw data, generate standardized event primitives, and output to risk identification, knowledge graph and early warning module. The present application can reduce the transmission pressure of raw data, improve the unified semantic expression ability of multi-source heterogeneous data, the fusion accuracy and the subsequent risk analysis efficiency, suitable for hoisting, high altitude, closed space and warehousing logistics and other smart engineering construction scenes.

The present disclosure relates to a headlamp system for a moving object and an operating method thereof, and more particularly, a headlamp system for a moving object can be provided, which includes a controller configured to generate light pattern information based on external environment information, and a light output device including a plurality of light sources and outputting a road projection image generated based on the light pattern information, wherein the controller calculates the light pattern information based on the external environment information and provides the light pattern information to the light output device, the light pattern information being data about an image of an update region that has changed compared to a previous road projection image.

The invention discloses a multi-master-device-oriented global bandwidth scheduling and priority control method and system, and the method comprises the steps: forming a complete end-to-end scheduling system through a software marking mechanism, network-on-chip real-time monitoring, initiating end adaptive adjustment, transmission path adjustment and a master memory controller cooperation strategy; and dynamic management of multi-service flow access behaviors is realized. In the system, a software layer configures initial strategies for different data streams; the main equipment end dynamically adjusts the priority of the sent request based on the real-time load; the network-on-chip carries out secondary adjustment on the priority based on the congestion state of the transmission path; the main memory controller maps the request to different queues for execution according to the final priority; and the system feedback layer collects the overall performance state and feeds back the overall performance state to the software layer so as to dynamically update the strategy. According to the method, software and hardware collaborative and adaptive global bandwidth management is realized, and the overall performance and service quality of the complex system on chip are remarkably improved.

This invention proposes a method, apparatus, equipment, and storage medium for simulation training of ship section construction. It achieves the construction of a high-performance virtual reality simulation environment, supporting collaborative interaction across multiple terminals such as VR, PC, and CAVE, significantly improving the system's adaptability and interactive immersion. Through a multi-user network synchronization architecture and workstation-level dynamic simulation scene construction, it realizes the realistic restoration of section construction process logic and operational synchronization, improving collaboration and training realism. Employing a multi-level detail model and dynamic mesh optimization technology based on view distance and bounding box overlap rate effectively reduces the rendering load of complex scenes and improves the system's real-time response performance. By establishing a multi-projection plane world coordinate system and real-time viewpoint matrix calculation, it achieves consistent rendering across multiple terminals. Combined with workstation cycle data synchronization and an operation scoring mechanism, it forms a quantifiable training evaluation system, thereby comprehensively improving the efficiency, safety, and scientific rigor of ship section construction training.

The invention discloses a fault detection method based on a fusion network, and relates to the technical field of fault detection of industrial equipment. The convolutional neural network model carries out local feature extraction and sliding window operation on time series data of equipment through a convolutional layer to capture a local mode and a change trend, and the Transform module helps the CNN model to capture global dependency relationship features in the time series data through a self-attention mechanism; the bidirectional gating recursion module considers forward and backward information of time series data at the same time and captures front and back dependency relationship characteristics, introduces the cross attention mechanism module to carry out characteristic fusion, and compresses the characteristics into fixed-length characteristics through adaptive average pooling mapping. The processed features are classified through a full connection layer, and prediction results of fault categories are output; according to the fault detection method based on the fusion network, the feature information of the industrial equipment can be automatically extracted, and the intelligence and the accuracy of the equipment are improved.

The present application relates to a kind of method and device for the distributed scene engine of vehicle, the method includes: obtaining the context information of vehicle;Current scene where vehicle is located is judged to be comprehensive scene or pure in-vehicle scene based on context information and predetermined rule set;When current scene is judged to be comprehensive scene, in-vehicle scene signal is generated for in-vehicle scene in comprehensive scene based on context information;And in-vehicle scene signal is sent to cloud server.

The invention discloses a network intrusion tracing method and system based on behavior analysis, and relates to the technical field of network security. The method comprises the following steps: S1, constructing an attack activity propagation graph and calculating an attack propagation coefficient between nodes; s2, according to behavior data and attack propagation coefficients of each node in the attack activity propagation graph, calculating behavior pulse factors of the nodes, and determining attack participation nodes; s3, calculating a time propagation coefficient of an attack path and constructing an attack time trajectory according to interaction time difference characteristics and behavior pulse factors among the nodes; s4, according to the flow rate, the attack propagation coefficient and the time propagation coefficient between the nodes, calculating a traceability flow overflow index, and generating a potential attack source set; and S5, calculating an attack intensity adjustment coefficient according to the traceability flow overflow index and the flow rate, and triggering a defense strategy response mechanism based on the attack intensity adjustment coefficient. Precise traceability and real-time defense are realized through multi-dimensional behavior analysis, and the network intrusion response efficiency is improved.

The invention discloses a real-time distributed actual battlefield scene simulation control system, and relates to the technical field of real-time distributed simulation, and the system comprises three core modules: 1, a distributed architecture construction module for constructing a deployment architecture combining a central node and an edge node, accessing a hardware terminal and formulating a unified data interaction specification, the system comprises a first module, a second module, a clock synchronization and load monitoring module, a task migration execution module and a third module, wherein the first module is used for guaranteeing node time consistency through a user-defined synchronization mechanism, collecting load data, calculating a dynamic association load value and setting a load threshold value in combination with a training scale; the migration cost is quantified, the task with the minimum migration cost is preferentially migrated, data transmission specifications and space-time consistency are ensured, actual combat training requirements of different scales are met, and system load balance, real-time response and scene high fidelity are achieved through the content of the method.

The application provides a user intention conversion processing method and device, computer equipment and a readable storage medium, and relates to the field of artificial intelligence. The method comprises the following steps: receiving a first session request of a user, wherein the first session request at least carries a call identifier and historical dialogue content; generating a structured session context based on the historical dialogue content, and generating a robot main reply result based on the structured session context; starting an asynchronous intention recognition process relative to the generated robot main reply result to perform intention recognition on several rounds of dialogue of the user, and obtaining an intention recognition result; if the intention recognition result is that the user has a conversion intention, writing the intention recognition result into a state cache, and associating the intention recognition result with the call identifier in the state cache; receiving a second session request of the user, wherein the second session request at least carries the call identifier; and when it is determined based on the call identifier that the intention recognition result is that the user has a conversion intention, performing a user conversion action.

The present application relates to a kind of multi-channel fast cooperative control method and system based on programmable event trigger mechanism, belong to chip testing field.The method aims at solving digital-analog cooperative test efficiency, timing is difficult to be strictly synchronized and the problem of long time consumption of channel switch operation under multi-station change scene.The present application is mainly in the configuration phase prestore event containing full-station channel enable information;Digital board card identifies event microinstruction in operating phase, and then by main control board card dispatches hardware synchronization signal line array, realizes digital-analog closed loop handshake;Analog board card carries out logic and operation with prestore information and real-time multi-station enable information, and automatically determines target physical channel.The present application eliminates computer interaction in execution process, realizes nanosecond level synchronization, and is mainly used to improve the efficiency and real-time response capability of multi-station chip testing.

The application discloses a mixed expert model word grouping scheduling method and electronic equipment, and relates to the technical field of artificial intelligence, which comprises the following steps: in response to the arrival of word units in the form of a stream to a preset buffer, obtaining the embedding vector features, routing history and arrival time corresponding to each word unit to splice into the target feature vector corresponding to each word unit, and the preset buffer is configured with a preset waiting time threshold and a capacity upper limit value; clustering and grouping each word unit according to the target feature vector to generate at least one task group; determining the priority of each task group based on the average waiting time of each word unit in each task group and the historical hot index; under the condition of meeting the limitation of the preset buffer, scheduling each task group according to the priority to enable the mixed expert model to perform operation on each word unit. The method of the application improves the utilization rate of experts and the system throughput while ensuring the upper bound of delay through short window buffering and dynamic clustering.

This invention relates to the field of protocol conversion and transmission technology, specifically an FPGA-based AFDX to Aurora protocol transmission and conversion system. The system includes: an AXI-AFDX module responsible for parsing and encapsulating standard AFDX frame signals, and separating AFDX control signals from data signals; an AFDX over Aurora module that achieves cross-clock domain synchronization by storing control and data signals separately in dual FIFOs, extracting frame groups in an orderly manner based on virtual link static priority scheduling and same-priority polling mechanisms, and encapsulating them into IP frame format with extended fields to ensure that critical AFDX control information is not lost during Aurora transmission; and an Aurora protocol module that performs bidirectional conversion between IP frame signals and Aurora protocol high-speed serial signals, adapting to external GT high-speed interfaces. This invention achieves reliable transmission to high-speed serial links.

The application discloses an unmanned ship path tracking control method based on 13-point adaptive PID anti-noise, relates to the path tracking technical field, and comprises the following steps: S1, a three-degree-of-freedom model of unmanned ship movement is constructed; S2, according to the three-degree-of-freedom model of unmanned ship movement, the speed of the unmanned ship machine is adjusted and the state is updated by using an adaptive PID controller and a 13P-ADAZND model constructed by 13-point ZTD format. The application adopts the 13-point ZTD format, the truncation error can reach seven orders, the numerical phase time delay is greatly reduced, compared with the low-order method, the application can reduce the deviation in high-precision operations such as turning and berthing, and higher path tracking precision is realized.

The invention discloses an intelligent safety protection system and method for electric power operation. The system comprises a multi-source fusion positioning and three-dimensional modeling module, a safety state perception and risk pre-judgment module, a hierarchical intervention control module, an edge calculation and digital twinning module, a 5G slice communication transmission module, a multi-scene adaptation module and a system data management and optimization module. The UWB, Beidou and visual reconstruction multi-source fusion positioning technology is adopted, dynamic three-dimensional modeling is combined for positioning deviation correction, the positioning limitation of a single technology in a complex environment is effectively solved, high-precision positioning is achieved, and accurate position support is provided for safety protection. According to the method, infrared thermal imaging, millimeter wave radar and a YOLOv8 optimization algorithm are fused, dominant safety violation behaviors can be identified, hidden misoperation risks can be pre-judged based on data such as equipment temperature and operation postures, and full-dimension safety state sensing is achieved.

The invention discloses a system and a method for adjusting a dynamic learning rate of TPM key agreement of a tree parity machine based on synchronous history. Aiming at the technical problems of low efficiency and insufficient stability caused by the fact that a fixed learning rate mechanism in traditional TPM key negotiation is difficult to adapt to dynamic change of a synchronization process, the system records a recent synchronization result by introducing an annular buffer area, and calculates and quantifies a synchronization progress index according to the recent synchronization result. And a negative correlation mapping function between the synchronization progress and the learning rate is constructed, and adaptive dynamic adjustment of the learning rate is realized. According to the method, convergence can be accelerated by adopting a large learning rate in the initial stage of synchronization, and stable and accurate synchronization can be ensured by adopting a small learning rate in the later stage of synchronization, so that the efficiency of key negotiation and the robustness of the system are remarkably improved on the premise of ensuring the safety.

The invention, which relates to the technical field of intelligent sensing, discloses a protection area multi-modal sensing fusion wild animal intelligent identification system comprising an intelligent sensing module used for receiving and obtaining a multi-modal data set; the fusion analysis module is used for processing the multi-modal data set, identifying all predation interaction events in a current sensing unit in a preset time window by adopting a target detection algorithm, generating a structured behavior semantic tag, and calculating an instant risk value of a generation event level of each predation interaction event so as to construct a food chain stability index; the first judgment module is used for executing first judgment to obtain a first evaluation result; aggregating the jth sensing area, calculating an average value of foraging efficiency coefficients FEC of all predation interaction events, and executing second judgment to obtain a second evaluation result; and the decision module is used for combining the first evaluation result and the second evaluation result to generate a corresponding strategy.

The invention discloses a big data intelligent recommendation method and system based on multi-scene dynamic adaptation. A dynamic user portrait is generated by constructing a three-dimensional scene feature space and fusing multi-source heterogeneous data; scene-feature weight adaptive adjustment is realized based on reinforcement learning, and a multi-scene model is trained in combination with a'shared-private 'neural network and transfer learning; and finally, dynamically allocating resources according to the scene priority to generate a recommendation result, and continuously optimizing through a feedback closed loop. The problems that a traditional recommendation system is insufficient in scene adaptation, poor in real-time performance and the like are solved, the recommendation precision and the resource utilization rate are remarkably improved, and the method is suitable for multiple fields such as e-commerce and content distribution.

This utility model relates to the field of cable tube stranding machine technology, and discloses a tube stranding machine for cable production, including a base, a tube stranding machine body mounted on the base, a stranding mechanism mounted on the tube stranding machine body, a cable body inserted through the stranding mechanism, a winding mechanism mounted on the tube stranding machine body, and a wire binding assembly mounted on the tube stranding machine body. The wire binding assembly includes a base frame fixedly connected to the base, a ring fixed on the base frame, a support plate fixed to the end of the ring, a telescopic rod fixed to the bottom surface of the support plate, and a clamping ring fixed to the end of the telescopic rod. This utility model solves the problem of incomplete binding caused by fixed specifications in traditional wire binding mechanisms, achieving a stable binding effect for cables of multiple specifications, and can apply balanced binding force to the cable from multiple directions, avoiding surface damage or structural deformation caused by excessive local stress, thus improving the appearance quality and structural integrity of the cable product.

The invention discloses a message scheduling system and method based on a low-delay message real-time distribution mechanism, and the system comprises an interface interaction assembly, a message receiving and transmitting assembly, a message scheduling management assembly, a message processing assembly, and a message storage assembly. The display module is used for displaying effective information for a user and providing a visual and convenient operation entry to realize information interaction between the user and the system; the message transceiving component is used for sending an instruction to the tested system based on the communication serial port and receiving feedback data, and is used for completing data communication between the low-delay message real-time distribution system and the tested system; the message scheduling management component is used for distributing and processing the received mass data, ensuring the integrity and accuracy of the received data, and improving the processing efficiency and response speed of the system through a message processing parallelization mechanism; the message processing component is used for extracting an effective part in the received original binary number and analyzing the effective part into a numerical value which can be intuitively read by a user; and the message storage component is used for automatically storing effective data and recording software operation logs.

The application discloses a data life cycle management method for psychological conversation of minors, an electronic device and a computer readable storage medium. The method implements differentiated storage and life cycle management on conversation data with different sensitivity levels and access frequencies by constructing a three-layer heterogeneous storage architecture composed of a memory storage unit, a document storage unit and a relational storage unit. The TTL mechanism of the memory storage unit is used to perform automatic recycling of short-term data, and the built-in TTL index of the document storage unit is used to perform automatic physical deletion of expired data according to the document creation time. The application also realizes zero loss of messages after service abnormal restart through the creation and recovery scanning mechanism of the in-transit mark key. The application solves the problems of high conversation context loading delay, manual data cleaning dependence, and high privacy leakage risk caused by mixed storage of conversation original text and metadata in the prior art, and provides a high-reliability data management scheme for the psychological health service scene of minors.

The invention discloses a multi-mode intelligent analyzer based on an FPGA, and the analyzer comprises a signal input unit which is used for carrying out the filtering processing of two paths of inputted analog signals; the analog-to-digital conversion unit is used for performing analog-to-digital conversion processing on each path of filtered analog signal to obtain a corresponding time domain signal; the wireless communication module is used for receiving a preset parameter threshold value set by a user; the FPGA main control unit comprises two parallel signal processing assembly lines which are used for carrying out frequency domain transformation on the received two paths of time domain signals in parallel to obtain frequency domain signals; the intelligent judgment module is used for performing parameter analysis on the frequency domain signal to obtain parameter characteristic data, and judging a preset parameter threshold value and the parameter characteristic data to obtain a judgment result; the display information fusion module is used for fusing the time domain signal, the frequency domain signal, the parameter characteristic data and the discrimination result to obtain a fused video data stream; and the display unit is used for displaying the fused video data stream and realizing multi-modal data display on the same screen.

This invention discloses a visual detection method for evaluating the standardization of drainage in filtration operations. By employing a state machine dynamic triggering mechanism, the optimal observation window is defined as the stable existence of the funnel and the absence of glass rod obstruction. The method accurately extracts and caches the single-channel binary mask of the three filter paper layers to avoid real-time segmentation noise. Then, based on the mask, the orientation of the three filter paper layers and the contact position between the glass rod and the filter paper layers are determined in real time. Relying on two-dimensional conditions, the method identifies when the glass rod is ready. Real-time interactive analysis and timed segmentation are implemented to achieve extremely high real-time response performance and reduce hardware requirements. Finally, the method calls the opposite-side drainage rules to adjudicate compliance, achieving full-process automation. This method ensures segmentation accuracy and judgment accuracy while reducing real-time computing power consumption, efficiently solving the core problems of existing technologies and providing reliable technical support for evaluating the standardization of experimental operations.

The invention relates to the technical field of sequence recommendation, in particular to a cross-domain recommendation method based on stream matching. The method comprises the following steps: constructing a training data set; determining a continuous velocity field for learning from the source domain preference distribution to the target domain preference distribution; constructing a deterministic linear conditional probability path connecting the first embedded representation and the second embedded representation by using stream matching; parameterizing the continuous velocity into a velocity field model to fit an optimal velocity field induced by the linear conditional probability path; constructing an alignment adaptation model to be trained; and adopting a decoupling optimization strategy to train the velocity field model and the alignment adaptation model to generate a cross-domain recommendation model. According to the cross-domain recommendation method based on stream matching, the preference semantic structure of the source domain can be efficiently kept, and the accuracy of cross-domain recommendation is improved.

The invention discloses a multi-modal sensor environment perception fusion prediction method based on a neural network, and relates to the technical field of computer-aided prediction. The method comprises the following steps: collecting multi-channel data and a calibration reference value of a sensor array; preprocessing and constructing a data set; constructing a fusion neural network including frequency domain modeling, a linear trend and a nonlinear residual branch; designing a joint loss function fusing data fitting and a physical prior regular term, and introducing an adaptive learning mechanism of regularization weight to carry out joint training; and finally, performing fusion prediction on real-time sensor data by using the trained model. According to the method, frequency domain global modeling and trend separation are unified through a three-branch explicit decomposition architecture, the precision, robustness, cross-working-condition consistency and real-time processing efficiency of multi-modal sensor fusion prediction in a complex industrial internet environment are effectively improved in combination with an adaptive training mechanism, and deployment on an edge platform with limited resources is facilitated.