27results about How to "Shorten the deployment cycle" patented technology

The invention discloses a parameter value estimation method for xPos relative position coding and a multi-level relative position coding method and system, relates to the technical field of deep learning, solves the technical problem that a Transform model is relatively weak in robustness and relatively low in accuracy when processing a long sequence, and is characterized in that an exponential decay factor is introduced into xPos relative position coding, so that the robustness of the xPos relative position coding is improved. According to the method, exponential decay is used for adjusting the encoding mode of query and key vectors, oscillation of high-frequency components is restrained, the capturing capacity of the model for long-distance dependence is enhanced, and the length extrapolation capacity of the model in a non-length sequence task is further improved. Meanwhile, xPos relative position coding is used for intra-segment coding and inter-segment coding of a text sequence, double-layer coding is achieved, hierarchical structure information in the text sequence is captured by combining intra-segment coding and inter-segment coding, long sequences which are not seen in the training period can be processed more effectively, and the length extrapolation capacity of the model is improved.

The invention relates to a digital twin-driven distributed warehousing system and method, and belongs to the technical field of intelligent logistics. The system comprises a sensing execution layer, a network transmission layer, an edge computing layer, a cloud platform layer and an application service layer. The method comprises the following steps: acquiring storage physical entity data through a perception execution layer; the data is transmitted to the edge computing layer and the cloud platform layer through the network transmission layer; in the cloud platform layer, a high-fidelity virtual warehouse is constructed and synchronized through a digital twin engine; using a multi-agent reinforcement learning algorithm to perform centralized training in a digital twin environment to generate a distributed cooperative scheduling strategy; and issuing the strategy to an edge computing layer and a sensing execution layer, and driving an autonomous mobile robot cluster to execute a cooperative task. According to the invention, step-type improvement of the warehousing system in flexibility, intelligence and cooperative capability is realized, and the problems of rigid bottleneck, non-optimal decision and low cooperative efficiency of a traditional system are solved.

The invention discloses a pull-push continuum robot, and belongs to the technical field of continuum robots. The robot comprises a tendon pulling module, a middle module and a structural backbone pushing module. The middle module is arranged between the tendon pulling module and the structural backbone pushing module; the tendon pulling module transmits tendon driving force to the structure backbone pushing module through the middle module, and continuous robot structure deformation is completed. Tendon driving and rod driving are naturally fused, a complex spatial configuration can be achieved through at least one motor, the tendon driving rigidity is improved, and the robot control difficulty is reduced.

The application relates to the technical field of agricultural parameter processing, and particularly discloses a facility agricultural microclimate collaborative parameter optimization method and system based on PlantDGP, which comprises the following steps: an environment sensing platform maps environment data according to a preset spatial unit set to form a spatial grid state; an environment operation port loads a PlantDGP collaborative control instruction set and completes syntax verification and semantic verification, and after the verification is passed, a collaborative rule description section is compiled into a collaborative control task graph; then, a window instance is generated according to a response window description section, and a spatial grid state is compared with an environment target description section to form a collaborative control trigger event within the valid period of the window instance; after receiving the trigger event, the environment operation port generates a candidate collaborative action package set according to the collaborative control task graph and performs a degradation trigger judgment; finally, a multi-target evaluation is performed on the action package that passes the judgment to select an execution plan instance package, which is then sent to a device controller for execution to complete collaborative parameter optimization.

The invention discloses a bridge vehicle load sensing method and system based on an encoder-decoder, and the method comprises the steps: building a variational inference model based on the encoder-decoder, constraining the shape of a hidden feature space in the middle of the encoder-decoder to be a discrete form of a bridge floor, an encoder part in the self-encoder is regarded as a variational inference model of a bridge deck load, and a decoder part is regarded as a constraint part of a training encoder network; building a finite element model and loading an actually measured vehicle load sequence to obtain a response data set of vehicle load and structure dynamic deformation, and taking the response data set as a supervised data set; structural deformation data of real operation of the bridge is collected as a fine tuning data set, and the result of the pre-training stage is docked for a subsequent fine tuning stage; carrying out pre-training and fine tuning on the network model; and inputting real-time bridge structure deformation to deduce the vehicle load in the characteristic space. The problem that real-time distribution of bridge deck vehicle loads cannot be obtained through a traditional method is solved, and the model can be rapidly deployed.

Embodiments of the present application disclose a robot path arrangement method, a robot, a computer device and a computer readable storage medium. The method comprises the following steps: performing standardization processing on an acquired scene map to obtain map data, the map data being composed of at least one structured point; matching a corresponding scene material for each structured point based on the map data; and arranging a work flow according to the scene material and the structured point, and issuing the work flow to the robot. Therefore, the present application can automatically generate a work flow based on a map and matching materials, reduce additional personnel operations, and realize end-to-end automation of a closed loop from map analysis to work flow landing.

The invention discloses a terahertz metamaterial sensor reverse design method based on transfer learning and reinforcement learning, and belongs to the technical field of terahertz metamaterial sensor design. In order to solve the technical problems that an existing terahertz metamaterial sensor is low in reverse design efficiency and insufficient in generalization ability, and structural parameters meeting a preset performance target are difficult to rapidly obtain, a hybrid deep learning model is firstly constructed and pre-trained, and then a generalization prediction model is obtained through an adversarial transfer learning mechanism. Then, a reinforcement learning optimization environment is established, finally, a depth deterministic strategy gradient algorithm is utilized, a generalization prediction model is used as an environment simulator for training, optimal metamaterial structure parameters are output, and reverse design of the terahertz metamaterial sensor can be efficiently achieved. The method is suitable for quickly acquiring sensor structure parameters meeting the requirements of a target frequency band, an absorptivity peak value and a quality factor.

This invention relates to a regional generalization method for an adaptive graph spatiotemporal prediction model for wind power clusters, belonging to the field of wind power prediction and new energy power system dispatching technology. The technical solution is as follows: first, a regional generalization method for the adaptive graph spatiotemporal prediction model of wind power clusters is used to solve the adaptability and accuracy problems of cross-regional prediction; then, the transparency and credibility of prediction decisions are improved by relying on the model structure interpretability analysis method. This includes the following two steps: first, achieving effective generalization prediction; second, endowing the prediction with interpretability. The beneficial effects of this invention are: significantly improved generalization performance, reducing computational costs and deployment cycle; optimized prediction accuracy, enabling the adaptive graph spatiotemporal prediction model of wind power clusters to more accurately capture the spatiotemporal correlation between wind farms, reducing short-term wind power prediction errors by 10%-20%, and providing more reliable data support for grid dispatching; and enhanced interpretability, meeting the requirements of the dispatching system for model auditability and reducing operation and maintenance decision risks.

The utility model belongs to the technical field of sorting equipment, and particularly relates to modularized high-speed sorting equipment which comprises a rack, a conveying mechanism, a first optical sorting machine, a first discharging unit, a second optical sorting machine, a second discharging unit and a third discharging belt conveyor. The two high-speed optical selectors, the matched storage covers, the matched discharging belt conveyor and the matched air injection device are integrated in the same rack module to form a compact sorting unit, so that the precision error caused by on-site scattered installation is avoided, and the equipment deployment period is greatly shortened; the technical problems that traditional multi-stage sorting equipment is large in size, difficult to disassemble and assemble and high in maintenance cost are solved.

The invention belongs to but not limited to the technical field of power distribution systems, and particularly relates to a power module frame structure system and method.The power module is mainly composed of a low-voltage cabinet, a UPS, a frame, a mesh plate and a cable, the whole frame of the power module is composed of square steel pipes and steel plates, and groove welding is adopted for welding between all stand columns and a bottom frame and welding between the stand columns and secondary beams. Welding beads are subjected to full-length welding treatment, and the welding strength is guaranteed. A Q235B square steel tube of 150 * 150 * 5mm is selected as a basic base, the thickness of the UPS is different from that of the low-voltage switch cabinet, and a rear base of the UPS cabinet is plugged by a steel plate of 4 mm. 100 * 100 * 5 mm Q235B square steel tubes are selected as the stand columns of the frame, and square steel tubes of the same specification are also adopted between the two stand columns to play a role in obliquely supporting the stand columns. A wire binding beam is additionally arranged on the upper-layer cable cabling rack aiming at the wire outlet position, input and output cables are laid in a layered mode, the span is controlled within 800 mm, and a reinforcing plate is additionally arranged between the two layers of cabling racks. Telescopic lifting lugs are embedded in the frame base so that the power module can be carried and hoisted conveniently.

This invention discloses a few-sample inductor core surface defect detection method based on model interaction, belonging to the fields of machine vision and industrial defect detection technology. The method includes the following steps: S1 Data acquisition and image preprocessing, constructing normal samples, labeled samples, and unlabeled samples; S2 Constructing an unsupervised statistical model based on statistical learning; S3 Constructing a supervised semantic segmentation model; S4 Simultaneously inputting the inductor core image to be detected into both the unsupervised statistical model and the supervised semantic segmentation model for processing, generating segmentation results; S5 Quantifying the differences in detection results; S6 Updating the parameters of the unsupervised statistical model; S7 Generating pseudo-labels based on the unsupervised statistical model; S8 Updating the weights of the supervised semantic segmentation model; S9 Using batches of processed images to be detected, inputting them into the updated unsupervised statistical model and the supervised semantic segmentation model for detection, obtaining detection results, and analyzing and calculating system performance indicators.

The utility model discloses an inspection robot rapid deployment support platform which comprises a carrier plate, an embedded computer and a voltage stabilization module are fixedly arranged on the carrier plate, a plurality of data interfaces are formed in the embedded computer, and the voltage stabilization module is electrically connected with the embedded computer; the embedded computer further comprises a protection shell, the protection shell is connected with the carrier plate and arranged above the embedded computer and the voltage stabilizing module in a covering mode, the protection shell surrounds the upper portions and the side portions of the embedded computer and the voltage stabilizing module, a first opening is formed in the protection shell, and the data interface is exposed through the first opening for data connection. The utility model further discloses an inspection robot adopting the rapid deployment support platform. The rapid deployment support platform integrates automatic environment sensing, self-adaptive algorithm loading and cloud collaborative verification functions, the deployment period is greatly shortened, meanwhile, rapid adaptation of most mainstream robot models is supported, and rapid deployment is achieved.

The application provides a battery health state evaluation method, system, computer device and medium, and belongs to the technical field of battery health management and life prediction. The method comprises the following steps: performing full-life cycle cyclic charging and discharging test on a reference battery, collecting data samples at fixed sampling intervals, and extracting a fusion feature vector of the reference battery sample; training a reference prediction model by using the reference data sample; collecting a fusion feature vector of a target battery to be tested, calculating the distribution distance of the features of the target battery and the reference battery by using an MMD algorithm, combining the prediction loss of the reference battery to establish a double-loss function, taking the minimization of the total loss as the target, iteratively adjusting the reference model parameters by using back propagation, and obtaining an optimized model suitable for the target battery; and the SOH value of the target battery can be accurately predicted based on the optimized model. The method greatly reduces the data collection cost and model deployment period in the migration stage, and realizes the migration of the prediction model in different battery devices at a low cost and high precision.

The application discloses a kind of multi-robot intelligent manufacturing cooperation systems, it is related to industrial robot control and intelligent manufacturing technical field, solve the technical problems of the heterogeneous compatibility difficult of prior art multi-robot system, dynamic response delay is high, collaborative obstacle avoidance precision is insufficient, lack of quality closed-loop management and control and weak fault-tolerant capability.This system uses cloud edge end distributed four-layer architecture, including edge collaborative control layer, multi-robot cluster execution layer, whole-process sensing layer and digital twin closed-loop optimization layer, supporting TSN+5G double-redundancy communication bus module.The application realizes the plug-and-play adaptation of heterogeneous robots, millisecond-level dynamic task scheduling and fault takeover, high-precision collaborative motion obstacle avoidance, and quality closed-loop optimization of whole-process processing, and can widely adapt to the flexible production needs of discrete manufacturing scene, greatly improve the efficiency of production line operation, processing yield and continuous operation stability.

The present invention provides a method for managing safety configuration in an adaptive environment construction phase, including: extracting construction environment features and converting them into feature vectors, and identifying construction environment categories based on the feature vectors based on a pre-trained environment recognition model; establishing a configuration template library, the configuration template library including configuration templates under different construction environments, each configuration template containing specific configuration items and parameters under the construction environment; selecting a matching configuration template from the template library based on an adaptation engine according to the construction environment recognition result, and loading it into memory; dynamically parsing the loaded configuration template, and adjusting configuration items and parameters according to construction environment characteristics or safety specification requirements; injecting the adjusted configuration parameters into a connector or application in real time through reflection or configuration management API to complete configuration update. The method of the present invention can realize automation and intelligence of configuration management, and improve the efficiency and quality of software deployment and maintenance.

The invention discloses a bearing fault diagnosis method for a new working condition limited data scene and a related device, belongs to the technical field of bearing diagnosis, and provides a double-layer fractional order differential enhancement mechanism for solving the problems of scarcity of industrial field fault data and variable working conditions. Capturing fine time sequence fluctuation of the bearing vibration signal by using a multi-order fractional order differential operator; in the feature layer, an adaptive distributed order fractional order feature extraction mechanism is designed, and the dynamic evolution law of deep features is described by using the non-local correlation of fractional calculus. The overall framework depends on a meta-learning normal form, and the model obtains efficient initialization priori through meta-training on a known working condition task set. According to the method, the diagnosis precision and robustness under the condition of extremely few samples are remarkably improved, and the problem of rapid diagnosis deployment of the rotating machinery under complex and changeable working conditions can be effectively solved.

The invention discloses an air-liquid homologous prefabricated data center cooling system, belongs to the technical field of data center infrastructures, and aims to solve the technical problems of how to realize cooperative work of air cooling and liquid cooling, and how to carry out efficient cooling of cooling capacity gradient utilization according to cooling capacity requirements so as to realize optimization of overall energy efficiency. According to the technical scheme, the system structurally comprises a pipe network square cabin, an IT square cabin and an outdoor cold source, the outdoor cold source provides cooling water through a dry-wet combined closed cooling tower, the water outlet end of the dry-wet combined closed cooling tower is connected with an outdoor water supply ring network, and water is supplied to the IT square cabin through the water supply ring network; the pipe network square cabin is connected with the water outlet end of the outdoor water supply ring network and used for conveying cooling water. Wherein the IT square cabin comprises a container body, a micro module and a refrigeration pipe network, the micro module and the refrigeration pipe network are arranged in the container body, the micro module is integrated with a cold plate type liquid cooling server cabinet, a liquid cooling CDU and an air cooling precision air conditioner, and the cold plate type liquid cooling server cabinet is connected with the liquid cooling CDU through a secondary side pipeline.

The application discloses a kind of unmanned aerial vehicle cargo cabinet centering mechanism, belong to unmanned aerial vehicle cargo cabinet field.The device includes: frame structure, drive structure and connecting structure.Frame structure includes four groups of telescopic components one fixedly connected at head and tail;Telescopic component one includes sleeve one;Slipper one is slidably connected in sleeve one;The side wall of sleeve one and slipper one is respectively provided with sliding slot;Two groups of telescopic components one are arranged at intervals, and two groups of push rods are arranged between them;Two ends of two groups of push rods are respectively fixedly connected with sliding block;Sliding block is slidably connected in corresponding side sliding slot.The cooperation of frame structure, drive structure and connecting structure makes the unmanned aerial vehicle cargo cabinet centering mechanism can match different size unmanned aerial vehicle cargo cabinet docking warehouse platform, greatly reduces the customized design and manufacturing cost for different scenes, improves the modularization performance and flexibility of the unmanned aerial vehicle cargo cabinet centering mechanism.The whole is assembled quickly and conveniently, can shorten deployment cycle, improve installation efficiency.

This application discloses a method for rapid access and management of cross-brand robots based on RUAPL, belonging to the field of robot control technology. This method is applied to a multi-robot management platform and includes: receiving access requests from at least one target robot; extracting a skill list from the access requests to construct a skill semantic graph; abstracting the capabilities of the target robots based on the skill semantic graph to obtain a unified abstract description; performing RUAPL-based adaptation and conversion on the unified abstract description to generate downlink commands; verifying the downlink commands based on the target robot's digital twin sandbox; and, if the verification passes, sending downlink commands to the target robot through a control channel to configure the target robot. This method enables rapid access, semantic interoperability, secure control, and unified management of cross-brand robots in heterogeneous environments, reducing the operation and maintenance costs and integration complexity in multi-brand mixed scenarios.

The present invention discloses a JSON-based dynamic policy generation method, which belongs to the technical field of software design. The method comprises defining and storing business element pattern specifications, creating policies and configuring required business elements, constructing business logic rule condition expressions in JSON format, constructing business matching result descriptors in JSON format, compiling and generating semantic descriptors recognized by a policy execution engine, receiving requests, performing business logic calculations and returning results. The method solves the technical problems of dynamically generating and modifying policies in real time through semantic descriptors in JSON format, using a policy execution engine to receive requests and perform real-time calculations and return results, and improving the maintainability and response speed of business logic rules. The method provides a new method for dynamic policy configuration, which has zero intrusion into business systems, eliminates the waiting period from traditional code development to release, realizes rapid response to changes in business needs, and improves production efficiency.

The invention provides an intelligent low-altitude economic unmanned aerial vehicle flight implementation monitoring system which comprises a system formed by an airborne subsystem, an edge node and a cloud control platform in a cooperative mode and is in service linkage with a digital twin airspace. Wherein the airborne subsystem comprises a flight control and task calculation unit, a communication multilink module, an equipment health monitoring module and a load interface module; the edge node comprises a machine nest and an edge gateway, and has the functions of taking off, landing, energy supply, data aggregation and evidence anchoring; the cloud control platform comprises task arrangement and scheduling, compliance semantic compilation, airspace risk prediction, monitoring and intervention implementation, evidence management and an open interface; the problems that the implementation process of an existing monitoring system is invisible and cannot be verified, sensing fragmentation is achieved, algorithm splitting is achieved, and multi-task and multi-machine cooperation is difficult are solved.

This invention discloses a barcode defect detection method and system based on visual image generation and screening, belonging to the field of image processing technology. The method includes: preprocessing the barcode image to be detected to generate a standard barcode image; inputting the standard barcode image into an improved convolutional autoencoder (BarCode-CAE) model to generate a reconstructed image; calculating a composite anomaly score between the standard image and the reconstructed image; comparing the composite anomaly score with a dynamic threshold T obtained through statistical modeling of normal samples to classify the image into high-confidence good products, high-confidence bad products, or undetermined samples; for undetermined samples, activating a defect detection module based on traditional image processing algorithms for secondary detection to finally determine its attributes. This invention achieves high sensitivity and robustness in detecting subtle barcode defects, effectively balancing the false negative rate and the high false positive rate.

This utility model relates to the field of containerized intelligent refrigeration room technology, specifically a containerized intelligent refrigeration room, including a container body, an intelligent refrigeration system, an equipment integration compartment, an intelligent control system, high-efficiency heat dissipation components, and a quick-connect structure. The integrated container design shortens the deployment cycle of the room from the traditional 30-60 days to 2-3 days, meeting the needs of emergency response and rapid deployment. The zoned evaporators and independent electronic expansion valves work together to ensure that the temperature difference between different areas inside the container is ≤2℃, improving the temperature control accuracy by 60% compared to traditional room designs. Intelligent management reduces the frequency of manual inspections, and energy consumption monitoring and automatic adjustment reduce refrigeration energy consumption by 15%-20%. Fault response time is shortened to less than 5 minutes. Directional airflow circulation reduces the temperature inside the equipment integration compartment by 8-12℃, improves the operating efficiency of the refrigeration system compressor by 10%, and extends the lifespan of components by 2-3 years.

This invention discloses a method for generating and scheduling management agents based on a large model and dynamic workflow. The method includes: analyzing business requirements through a large model, automatically generating and instantiating specialized management agents according to capability matching; dynamically generating workflows based on a rule engine and node selection probability, and adjusting the process structure in real time according to the execution status; employing a centralized multi-agent scheduling algorithm to decompose tasks into atomic tasks, allocating and adapting agents according to comprehensive scheduling to achieve collaborative scheduling and load balancing; and establishing an execution feedback and optimization mechanism, collecting logs and analyzing them using a large model to update rule weights and scheduling parameters. This invention can automatically generate management agents, shorten deployment cycles, reduce costs, improve process flexibility and system stability, form a self-optimizing closed loop, and significantly improve the efficiency of multi-agent collaboration and the level of system intelligence.

The application relates to a tool generation method and device, computer equipment and a storage medium, which comprises the following steps: generating a key element set based on key information in an interface description text; generating an interface model by performing data generation on the key element set based on a rule reasoning algorithm; constructing a calling sequence diagram based on the interface model, and optimizing the calling sequence diagram to generate a connection logic framework; adjusting abnormal processing logic in the connection logic framework to generate an enhanced logic structure; generating a construction script based on the enhanced logic structure; verifying the construction script based on a verification strategy to obtain a verification result, and generating an executable connection tool based on the verification result and the construction script; performing simulation test processing based on the connection tool to obtain test data, and determining a connection configuration corresponding to the connection tool based on the test data; and deploying the connection tool and the connection configuration to a target system. The application effectively improves the generation efficiency of the connection tool, and realizes rapid deployment of system integration.

The invention provides an intelligent early warning system for data quality and channel state of network-related equipment of a new energy field station, relates to the field of power system automation, and realizes non-intrusive data mirroring between the network-related equipment and a dispatching data network switch by deploying an optical splitter array on a physical layer to ensure that the system completely conforms to the security protection regulations of a power monitoring system. The system comprises a protocol analysis processing module which automatically identifies and analyzes multiple types of electric power communication protocols such as a 104 protocol, a PMU protocol, an information protection protocol and an electric quantity protocol; the intelligent early warning analysis module establishes an equipment dynamic flow baseline through time sequence feature learning, and constructs a channel health degree and data credibility two-dimensional evaluation model; and the multi-stage early warning notification module divides four early warning grades according to the abnormal severity, so that graded differentiated warning is realized. An optical power self-adaptive control technology is adopted to ensure stable mirror image data quality; a baseline dynamic updating mechanism is designed to avoid abnormal data pollution; and a full-link abnormity traceability function is supported.

The utility model relates to a wind and light storage hybrid energy supply system for an unmanned aerial vehicle airport on a tower, which comprises a tower-staying platform, the tower-staying platform is arranged on a rod frame of an electric power tower, and the unmanned aerial vehicle airport, a wind driven generator, a solar photovoltaic panel, a power generation controller and a storage battery are arranged on the tower-staying platform. The wind driven generator and the solar photovoltaic panel are connected with the power generation controller, the power generation controller is connected with the unmanned aerial vehicle airport for charging, the power generation controller is further connected with the storage battery for storing electric energy, and direct current output is achieved between the power generation controller and the unmanned aerial vehicle airport and between the power generation controller and the storage battery. The system can solve the problem of electricity taking of airport equipment on a power transmission and distribution tower, provides energy for an unmanned aerial vehicle airport, extends the applicable scene of the airport, and realizes unattended inspection operation.