17 results about "Incremental algorithm" patented technology

The invention relates to the technical field of communication production management, and discloses a communication equipment production intelligent management system based on machine learning. The system comprises a production data acquisition module, a feature engineering construction module, a dynamic clustering analysis module, an anomaly detection engine module and a production decision optimization module. The production data acquisition module acquires multi-source sensor data in real time and converts the multi-source sensor data into a standardized sequence with a unified timestamp; the feature engineering module extracts a time domain statistical feature, a frequency domain energy feature and an equipment state association feature to generate a high-dimensional feature vector set; the dynamic clustering module adopts an incremental algorithm to divide clusters online; the anomaly detection module establishes a multi-level Gaussian mixture model based on a clustering label, and quantifies an anomaly probability through a mahalanobis distance; and the production decision module integrates the results to generate an equipment maintenance priority sequence and a production takt adjustment instruction. According to the system, intelligent monitoring and dynamic optimization of the whole production process of the communication equipment are realized, and the real-time change requirement of a complex production environment is met.

The application discloses a passenger flow corridor generation method, device and equipment and a storage medium. The method comprises the following steps: performing vectorization region division on a target region to obtain planar vector data; obtaining OD data of the target region to form an OD data set; constructing a triangular mesh based on the planar vector data by using an incremental algorithm; constructing a directed traffic network based on the planar vector data and the triangular mesh; matching the start and end points in the OD data set with the nodes in the directed traffic network, identifying the travel paths of the OD pairs in the OD data set, calculating the traffic flow of the edges in the directed traffic network according to the travel paths, and generating directed traffic flow vector data; and performing hierarchical visualization processing on the traffic flow vector data according to the traffic flow size to generate directed passenger flow corridors of different travel modes in the target region. The application can efficiently process passenger flow corridor generation tasks of multiple cities or regions in batches, and realize multi-mode and direction-oriented passenger flow corridor identification.

The invention discloses a rice milling device and method capable of controlling machining precision, belongs to the technical field of artificial intelligence, and aims to solve the technical problem of how to convert a traditional rice milling process depending on experience into standardized and programmable precise machining. Comprising a feeding assembly, a whitening assembly, a bran discharging assembly, a rice discharging assembly and a controller. The controller, the camera, the gate driving mechanism, the whitening driving mechanism and the pressure driving mechanism are matched, the milling degree of rice is analyzed through an AI algorithm built in the controller on the basis of rice grain surface images collected by the camera, and rice milling parameters are obtained; and based on rice grinding parameters, the gate driving mechanism, the whitening driving mechanism and the pressure driving mechanism are dynamically adjusted through an incremental PID algorithm built in the controller.

This invention discloses a control method for multi-model photovoltaic module arrays, belonging to the field of intelligent photovoltaic technology. The method includes: S1, acquiring current electrical parameters; S2, electrical parameter difference and threshold comparison analysis: calculating the difference between the current electrical parameters and the predicted electrical parameters, and determining whether the difference is greater than a preset threshold. If it is greater, proceed to step S3; otherwise, proceed to step S4; S3, adjusting the power supply control quantity and time series array values, then proceeding to step S5; S4, rearranging the time series array; S5, predicting the next time series electrical parameters; S6, two-dimensional FFT: performing a two-dimensional Fourier transform on the time series matrix to obtain the time-spectrum array; S7, predicting the electrical parameter increment; S8, MPPT algorithm control. By integrating frequency domain analysis, probabilistic prediction, and distributed parallel control, environmental abrupt changes and module differences are transformed into computable time-series spectral features, achieving "predictive" tracking of MPPT, which can significantly improve power generation efficiency under complex illumination and module difference scenarios.

The invention provides a method for evaluating a signal quality index SQI in anesthesia equipment, which belongs to the field of anesthesia equipment monitoring and comprises the following steps: S1, performing data preprocessing on electroencephalogram data acquired in a database; s2, performing EEG signal quality judgment on the preprocessed EEG data, including obtaining impedance characteristics, signal spectrum entropy, total power in bandwidth and electrotome interference; s3, iteratively calculating a signal quality index SQI; and S4, when the impedance characteristic weight is less than 90, correcting the signal quality index SQI, reducing the impedance characteristic weight to 0.3, and increasing the electrotome interference weight to 0.6. According to the method, the SQI index oscillation caused by sudden change of a certain type of features can be limited by using two independent features in combination with a real-time correction incremental algorithm, so that the purpose of accurately evaluating the signal quality is achieved, and the influence of disturbance on index data is avoided.

The invention discloses a method for rapidly predicting the elastic-plastic mechanical property of a one-dimensional periodic spiral winding structure, and relates to the technical field of spiral winding structure mechanical property prediction. Comprising the following steps: S1, constructing a finite element model of a one-dimensional periodic spiral winding structure, performing linear elasticity analysis on the finite element model, and extracting a strain concentration tensor; s2, based on the extracted strain concentration tensor, clustering the finite element model by using a clustering algorithm to obtain a reduced-order model after partitioning, and calculating an interaction matrix representing a mechanical relationship between blocks in the reduced-order model; and S3, inputting the interaction matrix into an online incremental algorithm to obtain an elastic-plastic stress-strain curve of the one-dimensional periodic spiral winding structure, and realizing prediction of the elastic-plastic mechanical property of the one-dimensional periodic spiral winding structure. According to the method, the elastic-plastic mechanical property of the one-dimensional periodic spiral winding structure can be efficiently and accurately predicted, and the analysis efficiency and optimization efficiency of the one-dimensional periodic spiral winding structure are improved.

The invention relates to the technical field of computer information processing, discloses a power professional question and answer term dynamic updating method, and aims to solve the problem of semantic inconsistency of a question and answer system caused by updating lag of a traditional static term library. The method comprises the following steps: constructing a power professional term knowledge base and establishing a term semantic association network; dynamically monitoring new terms through a web crawler and a dual-trigger mechanism; semantic verification and disambiguation are performed in combination with the language model and the knowledge graph; updating the term library and maintaining a change history by adopting a version-controlled incremental algorithm; and synchronously triggering incremental learning and model fine tuning of a semantic understanding module of the question-answering system to realize collaborative updating of terms and the question-answering system. According to the technical scheme, the term change of the power industry can be perceived in real time, the semantic consistency of the question-answering system is ensured, the adaptive capacity and service efficiency of the system are remarkably improved, and the manual maintenance cost is reduced.

According to the satellite tracking method provided by the invention, a high-frequency dynamic gradient increasing algorithm is combined with program tracking, so that factors such as ephemeris errors and atmospheric disturbance are compensated in real time. The correction direction is directly calculated by the gradient method, and compared with serial search of step tracking, the convergence speed is higher, the gradient method is particularly suitable for the rapid motion characteristic of an LEO satellite, and the average received signal strength can be remarkably improved; according to the method, a predictive switching strategy based on multi-target QoS evaluation is introduced, the next service satellite can be selected from the global optimum angle, and the short vision of a traditional strategy is avoided. By comprehensively balancing the link quality, the service duration and the switching time consumption, the scientificity of the switching decision is ensured, and the switching frequency and the total signal interruption time can be effectively reduced, so that the data throughput and the operation efficiency of the whole ground station are improved.

The invention discloses an employee demission risk estimation method based on vectorization semantic analysis. The method comprises the following steps: acquiring an employee communication record through an enterprise communication platform API and performing sensitive information desensitization processing; a pre-training language model BERT or text-embeding-ad-002 is adopted to generate a 768-dimensional vector representation, and the 768-dimensional vector representation is stored in a vector database; dynamic clustering is carried out on the employees by using cosine similarity calculation and an incremental K-means algorithm, and clustering precision is optimized by using a CLIQUE algorithm; when a demission application is received, calculating a demission risk density and dynamically setting a threshold according to a clustering scale to trigger early warning; a state space including clustering risk density, employee characteristics and historical intervention effects is constructed, and personalized intervention strategies including communication intervention, salary adjustment, occupational development and the like are generated by using a PPO reinforcement learning algorithm. According to the method, the hidden demission risk which is difficult to find by a traditional method can be identified, the risk identification is advanced by 3-6 months, the high-risk employee identification accuracy is improved by 40%, the retention success rate is improved by 35%, the employee privacy is protected, and the data compliance requirement is met.

The invention relates to the technical field of block chain data security, in particular to an efficient block chain data verification and encrypted transmission method. The method mainly comprises three core steps of data preprocessing, data verification and encrypted transmission. In the data preprocessing stage, original data are partitioned, and identifiers are generated; in the data verification stage, an extensible hierarchical verification structure and an incremental algorithm are utilized, and data integrity verification can be rapidly completed only through partial hash values; in the encryption transmission stage, the data is symmetrically encrypted through a dynamic session key, and the key is safely distributed by means of a block chain smart contract, so that the method effectively overcomes the defects of low efficiency and insufficient transmission security of a traditional verification mode, realizes efficient and safe transfer of data between block chain nodes, and improves the security of data transmission. And the method is suitable for various block chain application scenes with relatively high requirements on performance and security.

The utility model discloses a kind of based on non-continuous track following racing robot, to solve the problem of speed restriction on non-continuous track in prior art due to inaccurate perception, control delay and unstable power supply. The utility model includes robot car body, drive motor, power module, main control board and the perception module consisting of camera and infrared sensor module. Main control board passes through output time division synchronous signal, cooperates and triggers camera and infrared sensor module to gather data and carries out fusion;Based on Kalman filtering and incremental PID algorithm, generate smooth motor control instruction when track marker is interrupted;And through real-time monitoring power load, dynamically coordinate system power consumption. The utility model improves the tracking speed, stability and endurance of robot on non-continuous track significantly through the cooperative optimization of perception-control-power supply link, while reducing hardware cost.

The application provides a tripping line type millimeter wave radar identification terminal and a cloud end object model synchronization system, relates to the Internet of Things sensing technical field, and the identification terminal comprises: a millimeter wave radar module which is arranged at the entrance and exit of a to-be-identified area to generate bidirectional flow statistical data in real time; and an edge computing device which is connected with the millimeter wave radar module and is used for performing bidirectional independent incremental analysis processing according to the bidirectional flow statistical data to calculate the number of target objects in the to-be-identified area in real time. The application analyzes the tripping line radar flow statistical frame through the edge computing device, maintains the in-door estimation, the in-door entrance and exit cumulative counts and the session by adopting a double-axis independent incremental algorithm, synchronizes the flow statistical attributes and the Internet of Things cloud platform object model by using the MQTT protocol, and can subscribe to the attribute under the optional linkage control, and is suitable for the non-contact flow statistical and occupancy management scene.

The invention discloses a computer data center resource scheduling method and system based on artificial intelligence, and relates to the technical field of data center resource scheduling and storage management, and the method comprises the steps: obtaining a time sequence sample, calculating a physical access time interval through the time sequence sample, carrying out the normalization of the physical access time interval, and obtaining a normalized access time interval; constructing a log-likelihood function based on maximum likelihood estimation, calculating an iterative update step length by adopting a Newton-Raphson incremental algorithm in combination with a fractional order log-bound function, and iteratively updating a shape distribution parameter and a scale distribution parameter; inputting the normalized access time interval, the optimal shape distribution parameter and the optimal scale distribution parameter into a risk assessment model, calculating an instantaneous risk rate, converting the instantaneous risk rate into a target access probability accessed in the next unit time, and constructing an inequality constraint trigger by combining the node retention penalty cost of a target data block and the network migration physical cost; and distributing the target data block to the matched target node.

The application discloses a data protection method and system based on field splitting and distributed desensitization, and the method comprises the following steps: obtaining an authorized external application, splitting sensitive information into several independent fields, and sending the sensitive fields to corresponding desensitization service nodes respectively, each desensitization service node is independently set and does not interact with any data, the desensitization service node generates a local unique and irreversible mask identifier MID based on the complete randomness of the received field plaintext data by using an incremental algorithm, stores the mapping relationship between the MID and the plaintext data, returns the MID to the external application, the application database of the external system does not store any plaintext data of the sensitive field, and all administrators have no access right to the complete data set of the desensitization service node. The application can realize that even if the application database or all desensitization service nodes are attacked, the complete sensitive information cannot be reconstructed, and the data security is maximally protected on the basis of not increasing the computing power cost.

This invention discloses an intelligent control method for a photovoltaic grid-connected inverter, relating to the field of photovoltaic power generation technology. The method includes the following steps: system check, maximum power point tracking (MPPT), real-time measurement of the output voltage and current of the photovoltaic array in the photovoltaic system, obtaining ΔP(K) using an incremental algorithm, and then using the MPPT algorithm to obtain P based on ΔP(K). max (K) Grid-connected control connects the inverter to the grid, enabling intelligent power adjustment. After grid connection, the inverter output is adjusted through a power optimization algorithm to bring it closer to the theoretical maximum power point. Safety maintenance measures include an incremental algorithm that calculates power increments in real time and updates the maximum power point accordingly, improving the accuracy of tracking the maximum power point and ensuring that the photovoltaic system can operate close to the maximum power point. The power optimization algorithm then adjusts the inverter output to bring it closer to the theoretical maximum power point, achieving real-time optimization and improving the overall performance and efficiency of the photovoltaic system.

The application discloses a multi-camera stereo vision system with non-overlapping FOV and a global calibration method, and relates to the technical field of aircraft structure strength test.The application adopts an encoding mark array as a control point, reconstructs three-dimensional coordinates of all encoding marks through an incremental SfM algorithm, and unifies the three-dimensional coordinates to a global coordinate system established by an aircraft body coordinate system, that is, the application first reconstructs the encoding marks in the global coordinate system; then sets a left camera coordinate system in an M-group stereo camera pair as a local coordinate system, acquires two-dimensional pixel coordinates of the center of the encoding mark detected by the left camera, and acquires a rotation matrix and a translation vector when the local coordinate system is converted to the global coordinate system by using a PnP algorithm, so as to accurately establish a space connection between the M-group stereo camera pair and the global control point, that is, to accurately establish a space connection between the cameras with non-overlapping fields of view, thereby realizing high-precision full-field deformation measurement of the wing.

The application discloses a kind of based on the speed regulating method of ship main engine speed control system of incremental PID algorithm, first preset ship main engine speed: preset the corresponding value of each gear injection quantity and speed;Then determine the relationship between injection quantity and speed;Finally, the ship main engine speed is adjusted.The speed regulating method of ship main engine speed of the application adopts PID control algorithm, the differential element in PID control algorithm can judge the change trend of ship main engine speed according to the change speed of deviation, realize the role of advance adjustment to main engine speed;PID parameter self-setting mode is used to avoid the influence brought by different ship types and different loads, increase the applicability of the method;Meanwhile, the application sets the trigger condition of main engine speed adjustment, avoids that main engine speed adjustment is too frequent, eliminates the shock caused by frequent action;And according to the change relationship between speed and injection quantity, the state of propeller in seawater is judged.