1106 results about "Data preservation" patented technology

The invention discloses an electronic archive data security system and method, and relates to the technical field of electronic document management. The electronic archive data preservation method comprises the following steps: acquiring a plurality of storage files to be stored and file attribute data of each storage file in real time; analyzing the risk level value of each storage file according to the file attribute data of the storage files, and marking the risk level value; classifying and storing the plurality of marked storage files into an electronic archive database; after a plurality of storage files are stored in a classified mode, time period access data of each storage file are obtained in real time, and real-time analysis is carried out in combination with risk level values. File attribute data are obtained in real time, and risk level evaluation is carried out on the file attribute data, so that the files can be marked and stored in a classified mode in real time; according to the method, the high-risk file is ensured to accept stricter protection measures, the risk assessment index is updated in real time, and the encryption strength is dynamically adjusted according to the risk assessment index, so that the protection capability on sensitive information is remarkably improved.

The invention relates to a method for improving satellite data inversion soil salinity by using hyperspectral data of an unmanned aerial vehicle. The method comprises the following steps: collecting a soil sample in a test area, and calculating soil salt content; acquiring and preprocessing satellite remote sensing image data of a test area; storing data, outputting a single-band satellite image, and calculating a spectral index; obtaining and preprocessing hyperspectral image data of the unmanned aerial vehicle in the test area; reconstructing and upscaling unmanned aerial vehicle image data pixels to a spatial resolution of 10m; reconstructing the narrow-band spectral data of the unmanned aerial vehicle into broadband spectral data consistent with a satellite image, storing and outputting a single band, and calculating a spectral index; screening a soil salt content characteristic variable, and constructing a soil salt content inversion model by taking the screened characteristic variable as an independent variable and the actually measured soil salt content as a dependent variable; and fitting a function relationship between the upscaling unmanned aerial vehicle soil salt content inversion model and the satellite inversion model by adopting a numerical regression function method to realize correction of the satellite inversion model.

The invention discloses a comprehensive algorithm for distribution and scheduling of AGV (Automatic Guided Vehicle) carrying tasks. S1: an SCADA system; equipment communication, data acquisition and storage, protocol analysis and logic processing; s11, the communication protocol, the address and the communication rhythm of the equipment are configured through the system, data acquisition data are stored in ES (ElasticSearch), the data acquisition data are stored for one year and backed up regularly, and the system achieves service logic processing through an LUA script. The carrying efficiency of the AGV is integrally improved, and the execution time of carrying a single task by the AGV is shortened by 10%; the production capacity of a workshop is improved, and the carrying rate exceeds 96% under the condition of no interference; the workshop logistics cost is reduced, and the overall number of AGVs is reduced by 10%; and the first pass yield of material carrying is improved by 15%. The deep fusion data analysis capability enables the whole system to predict demands more accurately, avoid conflicts, optimize the overall production efficiency, optimize the production process, reduce waiting and resource waste, and improve the flexibility of path planning and the reliability of task completion.

The invention discloses a heterogeneous data processing method and device and a storage medium, and relates to the technical field of data processing.The method comprises the steps that to-be-processed heterogeneous data are obtained, field identifiers in the heterogeneous data are extracted, and then the field identifiers in the heterogeneous data are identified on the basis of a pre-trained semantic identification model; performing semantic matching on the field identifier and a target field in a preset database to generate a field-level mapping relationship, generating a step-by-step processing logic adaptive to a heterogeneous data format according to the field-level mapping relationship, executing verification operation on the step-by-step processing logic in a code sandbox, and performing verification operation on the step-by-step processing logic in the code sandbox. And after the verification is passed, storing the output target field data to a structure database. According to the heterogeneous data processing method and device, the extracted field identification is subjected to semantic matching, and the extracted target field is verified in combination with the code sandbox, so that the accuracy of heterogeneous data processing is improved.

The invention discloses a laser radar point cloud real-time three-dimensional target detection method suitable for an unmanned vehicle. The method comprises the following steps: preparing a target detection data set; model training; updating parameters; simulation data preparation; performing model reasoning and result evaluation based on simulation data; model deployment and real vehicle data preparation; performing model reasoning and data storage based on real vehicle data; uploading and marking the data; and incremental learning and parameter updating. The three-dimensional target detection model provided by the invention specifically comprises a preprocessing module; a backbone network module; a neck network module; a feature selection module; and a detection head module. According to the three-dimensional target detection model provided by the invention, the point cloud data is converted into various aerial view images, and richer spatial information is mined; a two-dimensional distance sensing sparse convolutional layer is adopted to perform feature extraction, and multi-view feature fusion is realized through a double-branch network; a multi-scale feature pyramid network and a feature selection module are combined, so that the detection precision of a long-distance target and a small target is effectively improved; by constructing an efficient neural network architecture and combining training, simulation testing and real data incremental learning strategies, the robustness, precision and adaptive capacity of a target detection model are effectively improved, and powerful perception support is provided for an unmanned driving system.

The invention discloses a decision support method for time-space dynamic data of Internet of Vehicles based on knowledge retrieval enhancement, which comprises the following steps: collecting perception data inside and outside a vehicle, and separating general data and private data; preprocessing the data, and compressing and encrypting the private data; performing data processing by adopting a data knowledge dual-drive mode and a large language model, and converting the data into structured data; constructing a super-relation knowledge graph, and carrying out vectorization processing based on an embedded vector model; performing node importance evaluation based on the super-relation knowledge graph, and performing data updating and increment; storing the private data locally, uploading the general data to a public database of the cloud, and defining a dynamic access range of the public database; and extracting a user questioning entity by using the local entity recognition model, matching data existing in the local and the cloud, obtaining background knowledge contents, and generating an enhanced retrieval result. According to the method, the problems of acquisition of decision support information and use safety of data in an Internet of Vehicles environment are solved.

The invention provides an unmanned aerial vehicle intelligent inspection method and system combining a three-dimensional point cloud modeling and flight path planning technology. The system comprises a data acquisition module, a data processing module, a platform uploading module, a three-dimensional point cloud modeling module, a flight path planning module and a task execution module. By combining the three-dimensional point cloud modeling and flight path planning technology, the whole process is automatic, a closed loop is formed from data acquisition to flight path generation, manual intervention is reduced, automation, refinement and intelligentization of unmanned aerial vehicle routing inspection are achieved, and the routing inspection efficiency and safety are remarkably improved. The method has the advantages that point cloud data is processed through a pre-constructed point cloud data processing model, the quality of the point cloud data is improved, sub-centimeter-level modeling is realized in combination with a BIM technology, the precision of three-dimensional point cloud modeling is high, track planning is based on real three-dimensional geometry, and fine inspection is supported. And through data storage of different file types, the transmission processing efficiency is improved, the space information of the three-dimensional point cloud is fully utilized in the route planning, and refined inspection is supported.

The present invention provides a method, system, device and medium for automatically cataloging RQD of drilling cores, which belongs to the field of underground engineering drilling engineering informatization. The system of the present invention includes an image acquisition module, an image preprocessing module, an RQD calculation module and a cataloging module. The present invention collects core box images in batches in a standardized manner, uses a semantic segmentation network and a projection transformation to correct the distortion of the core box images, and then uses semantic segmentation to mark core blocks of 10CM and above, sorts the recognition results and calculates the RQD; uploads the on-site collected images to the cloud, and integrates the core box image correction and RQD calculation algorithm in the cloud, and performs image data storage, RQD calculation and RQD database cataloging in the cloud to achieve digitalized and standardized data backup. The technical solution of the invention can efficiently and standardizedly digitize and store core information, reduce the cost of mine data storage, improve data utilization, and provide data basis for mine rock mass quality evaluation.

The invention relates to the technical field of computers, and discloses a metadata storage method and device, computer equipment and a storage medium. Obtaining the current proportion among the average flash memory erasing times of the hard disk, the average power-on time of the hard disk, the metadata variation quantity and the metadata snapshot quantity, and determining the times proportion among the average flash memory erasing times; according to the average power-on time, the average flash memory erasing frequency and the frequency ratio, adjusting the current ratio to obtain a target ratio; determining a metadata variation, and generating a metadata snapshot in the memory according to the target proportion, the metadata variation and the metadata; and flashing the metadata snapshot in the memory to a hard disk corresponding to the metadata in the metadata snapshot. The problems that when metadata are stored, the proportion of delta and base is difficult to adapt to the current state of the SSD, and the service life and the average power-on time of the SSD cannot reach the optimal state are solved.

The embodiment of the invention provides a file receiving and reorganizing method and device, equipment and a storage medium. The method comprises the following steps: acquiring archive data, identifying, analyzing and uniformly formatting the archive data, screening the archive data according to a preset rule, and acquiring preprocessed archive data; and extracting metadata in the preprocessed archive data, matching the metadata with a metadata template in the format library, and storing the preprocessed archive data in a reorganization library if no metadata is missing in the preprocessed archive data. And performing four-property detection on the reorganized archive data in the reorganized library. If the reorganized archive data passes the four-property detection, the reorganized archive data is stored in an archive library through distributed storage, and if the reorganized archive data does not pass the four-property detection, the position of an error in the reorganized archive data is recorded, and a repair suggestion is provided. In this way, the efficiency and quality of archive reorganization can be improved.

The invention relates to the technical field of cell preparation quality evaluation, in particular to a cell preparation whole-process quality intelligent evaluation system and a method thereof, and the system comprises a data acquisition module, a preprocessing module, a mining module, a modeling module, a report automatic generation module and a data storage module. The modeling module is used for constructing a cell product index evaluation system and a product traceability model and comprises a cell culture state evaluation model and an environment monitoring analysis module, and the environment monitoring analysis module is used for carrying out traceability analysis by using a time sequence convolutional neural network; the automatic report generation module intelligently evaluates the quality of the whole process and generates a report, the data storage module manages data and model storage, sharing and calling, the system constructs a cell quality evaluation system based on multi-dimensional data, integrates multiple dimensional indexes such as morphology, surface markers and functionality, realizes comprehensive evaluation, and improves the evaluation efficiency. And the one-sidedness of single index evaluation is avoided.

The invention discloses a trust evaluation data storage tracing method and system based on a block chain technology, and particularly relates to the technical field of data storage management, and the method comprises the following steps: recording trust evaluation data in a block form based on the distributed storage characteristics of a block chain, automatic management of data submission, verification, storage and updating is realized by using an intelligent contract, data authenticity and non-tampering are ensured, time stamps are added for data through a timestamp function, a time-ordered operation chain is formed, a data tracing function is provided, an authorized user is allowed to inquire a data source and historical operation records, and the user experience is improved. Based on data access characteristics, a storage area is dynamically adjusted, and efficient utilization and optimal management of storage resources are achieved; according to the method, the storage position of data can be dynamically adjusted, hot data are stored in a high-performance storage area, cold data are migrated to a low-cost storage area, the access efficiency of high-frequency data is guaranteed, and the storage cost is optimized.

The invention provides a large language model reasoning method and device, electronic equipment and a storage medium, and the method comprises the steps: searching a storage space for a historical high-storage-efficiency tensor corresponding to each stored request according to a mapping table under the condition that a current request is received, and converting the historical high-storage-efficiency tensor into a historical target tensor; executing a current round of reasoning through the large language model according to the current request, obtaining an intermediate tensor corresponding to each layer of the large language model in the current round of reasoning process, and obtaining a current target tensor required by attention layer calculation of the large language model in the intermediate tensor; searching a current high-storage-efficiency tensor as a precursor of the current target tensor, storing the current request and the corresponding current high-storage-efficiency tensor in a storage space as historical intermediate data, and updating the mapping table; and inputting the historical target tensor and the current target tensor into the attention layer for calculation to obtain the reasoning text corresponding to the current request, thereby effectively reducing redundant calculation.

The invention relates to the technical field of three-dimensional point cloud scanning, and discloses a three-dimensional laser scanner and a real-time point cloud scanning method and system, and the method comprises the following steps: setting a dot spacing, setting the rotating speed of a drive motor, and setting the transmitting frequency of a laser transmitting and receiving module; controlling the motor to rotate to drive the laser to emit laser; when laser starts to be emitted, the reading module collects angle parameters of the coded disc in real time and measures a laser flight time value through the laser flight time measuring module; combining the angle parameter, the laser flight time value and the laser reflection intensity value read in the last time of laser emission into a frame of data and storing the frame of data in an original data buffer area; reading frame data of the original data buffer area in real time, and calculating to obtain a three-dimensional coordinate; and taking the three-dimensional coordinates and the laser reflection intensity value as a frame of data and storing the frame of data in a result buffer area. The problem that the real-time application requirement is difficult to meet in the prior art is solved, and the method has the advantage of being capable of reducing the operand.

The invention provides an active sentry monitoring method and device, a vehicle, a medium and equipment. The active sentry monitoring method comprises the following steps: acquiring approaching information of a target object outside a target vehicle by adopting a digital key system carried by the target vehicle; if the target object is not the vehicle owner, starting a camera carried by the target vehicle to record an initial video; if the vibration sensor carried by the target vehicle collects a target vibration signal, storing video data corresponding to a target time period in the initial video; a digital key system is used for obtaining approaching information of a target object, if the target object is not a vehicle owner, a camera is started to record a video, so that active sentry monitoring is achieved, energy consumption of a sentry system is reduced on the premise that the monitoring effect is guaranteed, meanwhile, a vibration sensor is used for collecting a target vibration signal to judge whether an abnormal event occurs or not, and the monitoring accuracy is improved. And storing the video data in the target time period to store the video data before and after the occurrence of the abnormal event, thereby saving the storage space and reducing the memory requirement of the vehicle-mounted memory.

The invention relates to the field of bridge area integrated navigation, and discloses a bridge area low-altitude complex environment integrated navigation method which comprises the following steps: synchronously aligning multi-source sensor data; multi-source data are fused for autonomous positioning of the inspection unmanned aerial vehicle; performing data storage based on a preset storage rule; planning the flight path of the unmanned aerial vehicle based on a full-coverage path planning algorithm; after the flight path planning of the unmanned aerial vehicle is completed, the unmanned aerial vehicle starts to move and completes autonomous navigation based on the autonomous following path model of the inspection unmanned aerial vehicle; based on the autonomous obstacle avoidance technology of the inspection unmanned aerial vehicle and the visual image, autonomous obstacle avoidance of the inspection unmanned aerial vehicle is performed by fusing the autonomous obstacle avoidance technology of the laser radar. According to the invention, the three-dimensional laser radar inspection unmanned aerial vehicle and the visual image can be fused, and accurate navigation positioning is provided for the unmanned aerial vehicle.

The invention discloses a form and data visual display method and system based on a low-code platform, and belongs to the technical field of software development, and the method is implemented by the following steps of: visually creating a form, and providing a core component for creating the form based on a webpage draggable technology so as to create a complete form meeting business requirements; the drag configuration comprises form layout configuration and form item configuration; persisting the form, storing configuration data of the form in a database, materializing a table structure, and storing a JSON object of configuration of the form; setting a form data model; setting access authority; accessing by a form filling person; performing dynamic rendering; and storing the filling data of the form filling person. According to the method, a large amount of repeated work can be solved, more concerned services can be developed, the efficiency can be quickly improved, the research and development cost can be saved, and meanwhile flexibility is kept.

Disclosed are example embodiments of systems and methods for hybrid automatic repeat request (HARQ). An example method includes performing a first HARQ compression to reduce a number of HARQ bits, the HARQ compression comprising one of a unified HARQ compression and a block-wise HARQ compression. Optionally, the example method further includes performing a second HARQ compression to further reduce the number of HARQ bits to a compressed number of HARQ bits. One of the first HARQ compression and the second HARQ compression include the unified HARQ compression and another of the first HARQ compression and the second HARQ compression comprising the block-wise HARQ compression. The example method also includes saving the compressed HARQ data to a DDR storage.

The invention provides a visualization method and device for the combustion process of a circulating fluidized bed boiler, and the method comprises the steps: collecting the operation data of the circulating fluidized bed boiler in real time through a distributed control system at an upper computer, and storing the collected data in a CSV format; in the lower computer, a three-dimensional modeling tool is used for establishing a three-dimensional model of the solid boiler, and grid division is carried out; controlling the upper computer to generate boundary condition data of numerical simulation according to the acquired CSV data, transmitting the boundary condition data to the lower computer, and reading the latest boundary condition data by using the lower computer to perform real-time simulation calculation; performing post-processing on a simulation calculation result by utilizing the lower computer to obtain a picture and curve data, and transmitting the picture and curve data to the upper computer through a network; and the upper computer is controlled to display the visual results of the temperature field, the velocity field, the pressure field, the gas component concentration field and the ash combustible content in real time through a platform interface in the forms of static diagrams, dynamic diagrams and trend curves. According to the method, the combustion process of the circulating fluidized bed boiler can be rapidly visualized, the operation state of the boiler can be predicted, and guidance is provided for operation optimization and parameter adjustment.

The invention discloses a GNSS (Global Navigation Satellite System) precision satellite clock error correction calculation method based on state space representation. The method comprises the following steps of: 1, selecting a data segment of 10 epochs as a fitting object, thereby effectively reducing the calculation amount, improving the speed of real-time calculation and ensuring the high efficiency and accuracy of clock correction correction fitting; 2, deducting the average value of the respective system clock correction of the current epoch according to the clock correction data of each epoch; 3, recovering the satellite clock error after the broadcast time is delayed for 5 seconds by adopting the SSR data before and after the detection, and carrying out secondary difference comparison on the clock error data after the gross error detection and the original clock error data; 4, the real-time precision clock error is recovered through the generated SSR information; and step 5, after the data of the real-time data stream received by the BNC is stored, real-time imitated clock error estimation is carried out. Compared with the prior art, the method has the advantages that high-precision positioning can be kept, and meanwhile the stability and performance of a positioning system are further improved.

The invention belongs to the technical field of computer software, and particularly relates to an automatic statistical method and system based on software defect texts, comprising a preprocessing module used for importing information containing software defect text data into the system, preprocessing the software defect text data, removing useless symbols and redundant texts in the texts, and storing the useless symbols and redundant texts in the system; the preprocessing module is used for preprocessing a. Txt file and storing the preprocessed data into the. Txt file in a specified format, the extraction module is used for extracting keyword features from the preprocessed. Txt file, the keyword features comprise at least one of defect identification, defect type, defect level, repair suggestion, defect category, defect description, processing suggestion, reason analysis and rectification scheme, and the statistical analysis module is used for carrying out statistical analysis on the keyword features. And the statistical analysis module is used for performing statistical analysis on the extracted keyword features. According to the method, related information of software defects can be efficiently and accurately counted, manual participation in column-by-column counting is reduced, and the efficiency of defect detection and management and the accuracy and efficiency of a statistical result are improved.

The embodiment of the invention provides a data processing method and device, equipment, a storage medium and a program product. Comprising the steps of judging whether a transmission mode of an application layer for a to-be-written data write request is a direct read-write mode or not; if the mode is not the direct read-write mode, continuing to judge whether a client cache mechanism is started or not; if the client cache mechanism is started, continuing to judge whether the to-be-written data or metadata of a parent directory of the to-be-written data carry a cache acceleration mark or not; if the to-be-written data or the metadata of the parent directory of the to-be-written data carry a cache acceleration mark, obtaining the size of an idle memory block of a preset cache; when the size of the free memory block is greater than or equal to that of the to-be-written data and the to-be-written data is not overlapped with the data in the preset cache, storing the to-be-written data in a memory block pre-allocated by the preset cache; and traversing the memory block list of the preset cache, and storing the to-be-written data to the target storage. The method is used for achieving the effect of improving the data processing performance.

The invention discloses a same-type commodity identification method and system based on a large language model, and electronic equipment. The method comprises the following steps: receiving a same-type commodity identification request initiated by a user; obtaining commodity data; carrying out structured processing on the commodity data; configuring instruction prompt words and necessary analysis attributes; combined cue words are generated by combining the same-type commodity recognition request, the commodity data, the instruction cue words and the analysis attributes; calling a large language model by using the combined cue word to generate standard attribute structure commodity data; carrying out grouping and similarity judgment on the commodities; configuring calculation attributes; grouping and calculating the commodity data forming the standard attribute structure; judging the same type of commodities according to the calculation result and the calculation attribute; and storing the data of the commodities of the same type in a database. Compared with the prior art, the method has the advantages that the requirement for manual intervention can be lowered, the method adapts to complex and diversified data, the recognition accuracy is improved, and the processing efficiency is improved.

The invention discloses a CAD application system design scheme data management method and system, and relates to the computer cartographic technology, and the method comprises the steps: in a target design scheme creation stage, separately creating design model data and design view data of a target design scheme, and copying to obtain intermediate design model data and intermediate design view data; a design scheme data editing stage: loading part of design model data associated with editing data, executing editing, and storing edited update data to the intermediate design model data and the intermediate design view data after editing; and synchronizing the intermediate design model data and the intermediate design view data to the design model data and the design view data of the target design scheme in a design data storage and submission stage. According to the method, through the data middle layer, high consistency and integrity of design scheme data can be well guaranteed through two-stage submission.

The invention discloses a large-area photovoltaic land screening method, which comprises the following operation steps: S1, collecting and storing regional land resource space data, and collecting and storing large-scale DEM (Digital Elevation Model) data of a specified regional terrain; s2, inputting a specified photovoltaic land use policy, a selection tendency and sensitive area data, screening out land type data meeting specified photovoltaic land use policy conditions and the selection tendency, and rechecking and removing land parcels related to the sensitive area to obtain a preliminary range; s3, inputting topographic factors influencing photovoltaic land selection, carrying out topographic data processing, extracting corresponding data and carrying out vectorization processing; s4, mountain shadow overlay analysis and slope overlay analysis are sequentially carried out, and analyzed land pattern spots are obtained; and S5, carrying out concentration degree analysis and screening on the land pattern spots to obtain a large-area photovoltaic land preferential selection area. The photovoltaic land can be automatically and efficiently screened in a large area range, and the efficiency and accuracy of photovoltaic power station site selection are remarkably improved.

The application provides a weather load-based air conditioner energy control method and device and a storage medium, wherein the method comprises the following steps: saving PLC bottom layer instrument data, indoor temperature sensor data and public data to a cloud platform server; processing the above data, calculating first index data in units of predetermined working time, calculating an optimized load amount according to the first index data; selecting a date closest to a predicted load day comprehensive temperature T in historical operation data as a reference day, taking the load of the reference day as a reference day load; and adjusting air conditioner energy according to the optimized load amount through two preset optimization conditions and the reference day load. The application predicts future load change rules by mining the corresponding relationship among climate, indoor temperature and load change rules through historical data, so as to realize the purpose of precise energy supply and effectively avoid energy waste.

The embodiment of the invention provides an intelligent penetration testing method and device based on deep reinforcement learning, and the method comprises the steps: constructing a network information environment, setting basic parameters, initializing the network information environment, and constructing an act-critic structure with an independent strategy network and an independent value network; modeling the Markov decision process of the penetration test into a quaternion group lt; s, A, R, Pgt; a Markov model is constructed; constructing a priority playback experience pool; performing penetration testing based on a Markov model, calculating and outputting an attack action to a network information environment through a strategy network based on a current penetration state to obtain a next penetration state, rewarding the attack action through a value network, and storing data in a priority playback experience pool; and calculating and determining a mode of preferentially selecting sampling according to the priority of experience, repeating the penetration test operation of the Markov model until the function is converged, and outputting an optimal penetration strategy.

The invention relates to the field of machining and manufacturing, in particular to a method and system for detecting the size of a metal forge piece, and the method comprises the following steps: calibrating a three-dimensional modeling camera, and completing camera calibration, hand-eye calibration and turntable calibration in combination with an intelligent error feedback system and a multi-calibration-plate auxiliary system; standard part scanning and three-dimensional modeling, including auxiliary positioning of punctuation, scanning modeling, point cloud processing, point selection and feature extraction, and data storage in a database; scanning a test piece and performing three-dimensional modeling, wherein the process comprises three-dimensional scanning and data storage of the test piece in a normal-temperature or high-temperature environment, point cloud processing, multi-level point cloud registration, feature extraction and matching, distance calculation and judgment on whether the test piece is qualified or not; and outputting and storing the measurement data, and generating a test report to record the measurement data for inspection. According to the invention, high-precision size detection of the metal forge piece with a complex structure in a harsh environment is realized, the accuracy of a result is ensured, the batch detection efficiency and reliability are improved, and the method is suitable for industrial application scenes with high requirements.

The invention provides a steering gear autonomous zero calibration control system and a steering gear autonomous zero calibration control method. The steering gear autonomous zero calibration control system comprises a data acquisition module used for acquiring required bus message information; the vehicle mass calculation module is used for calculating the vehicle mass; the steering wheel angle calculation module is used for calculating a corresponding steering wheel sensor angle value; the tire pressure calculation module is used for receiving the tire pressure measured by the tire pressure sensor; the straight-line driving judgment module is used for judging whether the vehicle is in a straight-line driving state or not according to the wheel rotation angle obtained by the wheel rotation angle sensor in the steering controller; the zero angle control module is used for vehicle zero angle control calibration; and the data storage module is used for storing the data in the corresponding NVRAM after the vehicle learns new zero angle data. The problem that a steering wheel is not correct due to the fact that the electric power steering zero angle of the vehicle deviates from the zero position due to changes of the mass and the tire pressure of the vehicle is solved.

The invention provides a communication method and device, belongs to the technical field of communication, and is used for improving the reliability of ML model evaluation. According to the method, the data of different ML models can be stored in different data sets, and the reliability of ML model evaluation is improved by using the data in the data set corresponding to the ML model. For example, after an AnLF functional entity receives information of a first ML model, the AnLF functional entity can set to store related data of the first ML model in a first data set, namely, set a corresponding relation between the first ML model and the first data set, and send the corresponding relation to an MTLF functional entity by carrying the information of the first ML model and the information of the first data set in a first message; therefore, the MTLF functional entity can obtain the related data of the first ML model from the first data set so as to evaluate the first ML model based on the related data, thereby improving the reliability of evaluation of the first ML model.