60 results about "Square error" patented technology

The invention relates to the technical field of intelligent agriculture and computer vision, in particular to a small-sample high-density chicken counting framework based on a deep learning Mamba structure, which comprises a feature extraction network, a support-query enhancement module and a decoder. According to the method, a multi-scale feature extraction network based on a residual block (ResNet Block) is introduced, so that local detail information is effectively reserved; and then, a support-query enhancement module is constructed by introducing a Mamba structure, and context interaction between support features and query features is effectively enhanced by utilizing the long sequence modeling capability of linear complexity of the support-query enhancement module, so that the problems of individual overlapping and boundary fuzziness in a high-density chicken flock scene are solved. Experimental results on a PoultryCount real breeding data set show that the mean absolute error (MAE) and the root-mean-square error (RMSE) of the method are reduced compared with those of an existing method, and the chicken counting precision and generalization ability under the condition of a small number of labeled samples are remarkably improved.

The invention discloses a J-A model parameter identification method, system, equipment and medium based on RBF and an improved brownish bear algorithm, and belongs to the technical field of power system optimization, and the method comprises the steps: building a Jiles-Atherton hysteresis reverse model of a current transformer, determining a to-be-identified parameter vector, and building a model with a root-mean-square error between actually measured magnetic field intensity and simulated magnetic field intensity as a target function, training a radial basis function neural network model, expanding data through linear interpolation processing, obtaining a predicted magnetic induction intensity value, inputting an objective function and radial basis function prediction data into an improved brownish bear optimization algorithm, and iteratively optimizing model parameters through hierarchical population position updating and fitness evaluation until convergence conditions are met. And outputting an optimal parameter identification result. According to the method, high-precision and high-efficiency identification of hysteresis model parameters is realized, the generalization capability and robustness of the system are improved, and reliable technical support is provided for hysteresis characteristic analysis of a complex physical system.

Disclosed in the present invention are a fault diagnosis method and system for blade icing and blade mass imbalance of wind turbines. The method comprises the steps of: acquiring operation data and blade icing information of a plurality of wind turbines; labeling the operation data with an icing state label on the basis of the blade icing information, so as to obtain fault data; extracting fault features in the fault data, and ranking the fault features according to the degree of importance, so as to generate an optimal feature set; on the basis of a criterion of minimizing a squared error, selecting optimal features in the optimal feature set to generate an optimal decision tree; and performing classification on the basis of the optimal decision tree, so as to obtain a diagnosis test result including fault information. The present invention has the advantages of a high level of diagnostic accuracy, etc.

The invention discloses a geophysical prospecting signal denoising method based on VMD (variational mode decomposition) combined with an improved wavelet threshold function, and belongs to the technical field of mineral exploration geophysical prospecting signal process.The method includes the steps that firstly, a decomposition mode number K of a signal is determined through VMD in a self-adaptive mode, a plurality of IMFs (intrinsic mode components) are obtained, and then according to the frequency characteristic and noise distribution of each IMF component, a wavelet threshold function is obtained; and carrying out targeted noise suppression by adopting an improved wavelet threshold function containing an adjustment parameter alpha, and finally, carrying out linear superposition reconstruction on all the processed IMF components to obtain a de-noised geophysical prospecting signal. Experimental verification shows that compared with a traditional method, the method has the advantages that the signal-to-noise ratio and the correlation coefficient of the noisy geophysical prospecting signals can be effectively increased, the root-mean-square error can be effectively reduced, the inherent defects of the traditional method are overcome, the geologic features of the geophysical prospecting signals can be effectively reserved, the method is suitable for various mineral exploration scenes, and reliable data support is provided for anomaly recognition in mineral exploration.

The invention relates to the technical field of electromagnetic scattering characteristic analysis and modeling, in particular to a scattering center establishing and correcting method based on a geometric model and reference data. The method comprises the following steps: inputting a target geometric or grid file, and extracting geometric features of a plane, an edge and a curved surface to construct an initial scattering center model; inputting electromagnetic scattering reference data of the same target, solving a scattering center correction coefficient by adopting a pseudo-inverse or Tikhonov regularization method, and optimizing an initial model; and outputting an optimized scattering center model with angle and frequency expansibility. According to the method, the electromagnetic scattering characteristics of metal and coating type complex structure targets in single-station, double-station and full-polarization scenes can be uniformly represented, and the problems of insufficient model precision, incomplete types and high data dependence degree in the prior art are solved; the method achieves the excellent effects that the RCS root-mean-square error is smaller than 2dB and the high-resolution imaging structure similarity exceeds 90% in typical target modeling, and is suitable for the fields of target characteristic simulation and characteristic extraction and recognition.

The invention discloses a strain sensor layout optimization method based on proxy model assistance, and the method comprises the steps: carrying out the finite element simulation of underwater equipment, and constructing a simulation strain field data set of the surface of the underwater equipment; dividing the underwater equipment into different candidate areas, wherein candidate measurement points are distributed in each candidate area; converting a strain sensor layout problem into an optimization problem of determining a candidate measurement point from each candidate region, thereby constructing a design space of the optimization problem; modeling is carried out on the optimization problem; in the optimization model, using a root-mean-square error between a simulation strain field and an interpolation strain field obtained by interpolation reconstruction as a real fitness function; and solving the optimization problem based on an SO-I algorithm to obtain an optimal strain sensor layout scheme. According to the method, the problem of layout optimization of the complex-structure strain sensors of underwater equipment is solved, and maximization of a coverage area with the minimum number of strain sensors is realized.

The invention discloses a digital surface model (DSM)-guided hyperspectral and LiDAR combined unmixing method, relates to the field of multi-modal image processing, and aims to solve the problems of end member confusion and insufficient space structure maintenance caused by spectrum similarity in hyperspectral unmixing. According to the method, a hyperspectral image and LiDAR data of the same area are obtained, a LiDAR elevation map is expanded into a multiband profile through an attribute configuration file method, and a digital surface model (DSM) is generated. Constructing a spectrum and space double-branch auto-encoder, respectively extracting spectrum and space features and carrying out fusion mapping, obtaining an abundance matrix by using a normalized exponential function, and reconstructing a hyperspectral image; a double-branch adaptive mixed channel attention mechanism is designed in a spectrum branch, and a space attention mechanism is introduced in a space branch. In the training process, a self-defined loss function is formed by combining DSM-guided structure entropy regularization, spectral angular distance and root-mean-square error, and the space continuity and boundary retention of the abundance graph are improved. The method is suitable for hyperspectral unmixing and multi-modal remote sensing fine identification under complex terrains.

The invention discloses a pressurized water reactor steam generator liquid level control method and system, and belongs to the technical field of nuclear power station control. The method comprises the steps that a liquid level set value and a measured value of the steam generator are obtained, and deviation is calculated; inputting the set value, the measured value, the deviation and the bias constant into a preset BP neural network; the network dynamically outputs three coefficients of proportion, integral and differential through forward propagation to serve as real-time parameters of a PID (Proportion Integration Differentiation) controller; the PID controller calculates and outputs a water supply flow demand value, and a water supply valve is adjusted after conversion; and calculating a root-mean-square error of liquid level control as a loss function, and updating the weight of the neural network on line by using the new liquid level deviation through a back propagation algorithm to realize adaptive optimization of PID parameters. According to the invention, the self-learning ability of the BP neural network is combined with the reliability of PID control, the control quality and stability of the liquid level of the steam generator under various power levels are significantly improved, the system can be deployed in a power plant DCS through modular packaging, and the engineering practicability is high.

The invention relates to the technical field of electric power systems, and discloses an electric power measurement data management system and method based on an intelligent terminal and an RPA, and the method comprises the steps: obtaining clock synchronization topological structure data, an original timestamp sequence and measurement data of intelligent terminals of a whole network; extracting a clock synchronization dependency relationship of the terminal based on a graph structure analysis algorithm, and generating a directed dependency graph; inputting the timestamp sequence and the dependency graph into a causal discovery algorithm, and identifying a clock error propagation path and causal intensity; constructing a graph neural network, taking an intervention effect prediction matrix as node feature input, and predicting a whole network cascade correction effect by minimizing a weighted square error between a corrected inter-node timestamp difference value and a reference difference value; according to the invention, systematic elimination of clock drift cascade influence in a power system is realized, and time sequence consistency of whole network measurement data is ensured.

The invention discloses a flood simulation verification method based on multi-source data fusion, and belongs to the technical field of flood disaster simulation and risk assessment. The method comprises the following steps of collecting at least two types of multi-source data in insurance claim settlement, Internet of Things monitoring, remote sensing images and social media, extracting real submerging information after preprocessing, and constructing a comprehensive verification data set; simulating a submerging range and water depth by using a flood model, calculating a hit rate, a fitting statistical value, a deviation score and a root-mean-square error index, allocating weights to each verification data source, performing weighted fusion to obtain a comprehensive verification index, and quantitatively evaluating simulation accuracy; model parameters are iteratively optimized based on comprehensive verification indexes, and simulation precision is improved; and finally outputting a visual result, and applying the visual result to insurance claim settlement and disaster emergency management. According to the method, the problems of single data, narrow coverage and poor timeliness of a traditional verification method are solved, high-efficiency and comprehensive verification is realized, and the reliability and the practical value of the flood model are remarkably improved.

The invention relates to the technical field of artificial intelligence, and discloses a regional ocean current analysis method and device, and the method comprises the steps: obtaining the observation data of a regional ocean current, carrying out the standardization processing of the observation data, generating standardized ocean current data, and calculating the data point distribution density of the standardized ocean current data, so as to obtain the data point distribution density of the regional ocean current, the method comprises the steps of performing dynamic self-adaptive grid division on regional ocean current based on data point distribution density, generating a non-uniform grid structure, performing numerical interpolation calculation on the non-uniform grid structure based on standardized ocean current data, generating an initial interpolation result of the regional ocean current, performing algorithm correction on the initial interpolation result, generating a corrected interpolation result of the regional ocean current, and obtaining the corrected interpolation result of the regional ocean current. Performing root-mean-square error analysis on the corrected interpolation result to generate a final interpolation result of the regional ocean current; the method can solve the problem that the analysis conclusion deviates from reality due to the reduction capability of the limit interpolation result to the actual ocean current distribution condition and error accumulation.

The invention provides a method for quantifying tidal non-stationarity, and belongs to the technical field of natural science research based on nonlinear data analysis. Firstly, hour-by-hour tide observation data with the observation duration not less than one month in a target sea area are obtained; separating non-tidal low-frequency and high-frequency tidal changes in the data by using an empirical mode decomposition method, and extracting a tidal component; returning the tide component through a classical harmonic analysis CHA model to obtain CHA return tide; then calculating a correlation coefficient, a mean absolute error and a root-mean-square error of the tidal component and the CHA, and respectively dividing the mean absolute error and the root-mean-square error by an average value of the observation absolute values to obtain a normalized index; finally, (1, 0, 0) is used as a reference point, (CC, NMAE, NRMSE) is used as an observation point, and the tidal non-stationarity index NS is calculated according to a two-point distance formula. The device can be widely applied to global sea areas and is suitable for sea areas of different tide types and sea areas of different tidal ranges.

The application discloses a differential privacy gradient adaptive clipping method and device based on minimizing gradient square error, which comprises the following steps: inputting a batch of sample data into a classification model and calculating a gradient, and calculating a two norm based on the gradient; constructing a privacy histogram of the two norm based on a preset range, and updating the range, resampling data, calculating the two norm and reconstructing the privacy histogram when judging that the range of the privacy histogram is not appropriate; constructing a candidate set according to a current clipping threshold, calculating the gradient square error of each clipping threshold in the candidate set based on the finally constructed privacy histogram, and screening the optimal clipping threshold of the current round based on the principle of minimizing the gradient square error; when judging that the optimal clipping threshold of the current round is not appropriate, constructing a new candidate set according to the optimal clipping threshold, and iteratively re-screening; and calculating the privacy overhead under the optimal clipping threshold. The method and device not only guarantee the model performance, but also greatly reduce the computing power overhead of the computing device, and are suitable for low-computing-power devices.

The invention discloses an earthquake sand liquefaction deformation prediction method based on machine learning, and relates to the technical field of earthquake engineering and machine learning. Missing value processing, abnormal value detection, feature selection and standardization processing are carried out on the multi-source coupling seismic data set to obtain a seismic sand liquefaction deformation data set, a seismic sand liquefaction deformation prediction model is constructed according to the seismic sand liquefaction deformation data set, the root-mean-square error, the mean absolute error and the decision coefficient of the model are calculated, and the seismic sand liquefaction deformation prediction result is obtained. And according to corresponding threshold judgment, carrying out hyper-parameter adjustment on the model through Bayesian optimization or carrying out model packaging through Flask, processing the obtained new seismic sand liquefaction characteristic data, and inputting the processed data into the packaged model to obtain a seismic sand liquefaction deformation prediction result. The method can effectively predict earthquake sand liquefaction deformation, and provides support for earthquake disaster prevention and engineering design.

The invention discloses an intelligent foundation pile defect identification method and system based on KPCA-Light GBM-SADE, and the method comprises the steps: collecting and preprocessing the original data of low-strain reflected waves of a foundation pile, and obtaining the preprocessed waveform data; the method comprises the following steps of: performing adaptive decomposition on waveform data by adopting CEEMDAN to obtain IMF components, screening effective IMF components according to an energy ratio threshold, extracting composite multi-scale time-frequency features from a preprocessed waveform and the effective IMF components, and constructing an initial feature set; and screening feature subsets based on a mutual information criterion, and performing nonlinear dimensionality reduction on the feature subsets through KPCA to obtain a final feature set. SADE is utilized to adaptively and collaboratively optimize LightGBM hyper-parameters, and a recognition model is obtained through training with the foundation pile defect depth as a target. And inputting the final feature set of to-be-detected data into the model, outputting a defect position prediction value, and calculating a quantization confidence coefficient based on a verification set root-mean-square error to realize foundation pile defect identification.

The application discloses a pre-hospital care helicopter optimization deployment method fusing weighted K-means and IABC, first, patient data is acquired and the geographical position coordinates thereof are standardized pretreated, the patients are classified according to the injury degree and corresponding weight values are given, and a weighted data set is formed; subsequently, a weighted clustering square error and a silhouette coefficient are calculated, the elbow rule and the silhouette coefficient method are cooperatively decided, and the optimal helicopter deployment number k is determined; then, the weighted data set is taken as input, and the initial deployment position coordinates of the k helicopters are solved through a weighted K-means algorithm; then, a single-target fitness function integrated by three sub-functions including service coverage, rescue response time and economic cost is constructed, and a multi-target optimization problem is integrated into a single-target optimization problem; finally, an improved artificial bee colony (IABC) algorithm with a directional learning mechanism is adopted, the initial deployment position of the helicopter is taken as an initialization population for iterative updating, and the final pre-hospital care helicopter deployment coordinates are output.

The invention relates to the technical field of image detection, and discloses a combined template production quality detection method and system, and the method comprises the steps: collecting a side image through employing an RGB industrial camera in combination with multi-angle directional illumination and a signal-to-noise ratio self-feedback mechanism; executing channel selection, edge enhancement and dynamic narrowband ROI extraction; identifying interlayer boundary line segments by adopting a direction angle filtering and line segment merging mechanism; establishing an offset vector function; and outputting the offset and the root-mean-square error result. Compared with the prior art, the technical problem that accurate identification and quantitative evaluation of dislocation between the side surfaces of the plywood cannot be realized under the actual production conditions that slight dislocation occurs between the plywood layers and the contrast ratio of edge textures is insufficient in a detection mode for judging the lamination quality of the multilayer plywood by depending on manual vision or a simple image threshold method in the prior art is solved. Due to the fact that the structured boundary extraction and quantization vector modeling mechanism is adopted, the detection accuracy and the automation degree of the slight dislocation defect are improved.

The invention relates to the technical field of vibration recognition, and provides a method and a device for recognizing the vibration continuous increase or decrease trend of large rotating equipment. The method comprises the following steps: acquiring monitoring data of a vibration measuring point of to-be-monitored large rotating equipment, wherein the monitoring data comprises a vibration value corresponding to each data acquisition moment acquired based on the vibration measuring point in a data acquisition period; performing linear regression fitting on the monitoring data to generate a regression equation; according to the monitoring data and the regression equation, the increment and the root-mean-square error are calculated respectively; and comparing the increment with a product of the sample capacity of the monitoring data and a preset growth coefficient, and comparing a root-mean-square error with a preset root-mean-square error reference value so as to perform continuous increase and decrease trend identification on the monitoring data and generate a trend identification result. According to the method, the monitoring burden of operators can be obviously reduced, problems can be found in time in the early stage of faults, and time is bought for judgment and disposal of the operators.

The invention discloses a differential dynamic microscopic image processing method based on a root-mean-square error. The differential dynamic microscopic image processing method comprises the following steps: a sequence image acquisition step: acquiring a time sequence microscopic image; a sequence image processing step: obtaining a difference image data set; performing one-dimensional Fourier transform in two directions on the differential image data set in sequence to obtain two-dimensional transform result data, and calculating to obtain Fourier amplitude spectrum image data; an obtained data processing step: carrying out dimension reduction processing on the Fourier amplitude spectrum image data; introducing structure function fitting, calculating a root-mean-square error between a fitting curve of the data after dimension reduction processing and the Fourier amplitude spectrum image data, and selecting a data truncation point according to the root-mean-square error; and an intermediate scattering function is introduced, and dynamic information of Brown particles is obtained based on the relationship between the data after selection of the truncation point and the diffusion coefficient. According to the method, the data truncation point is determined through the root-mean-square error, so that the processing time is remarkably shortened while the measurement precision is ensured.

This provides a technique for calculating updated values ​​that include the effects of disturbances. [Solution] One embodiment is a ship's automatic steering system that outputs a command rudder angle using hull parameters based on a reference bearing and a bow bearing, and comprises an identification calculation unit that calculates an identification value by adjusting the hull parameters such that an evaluation quantity, which is the sum of squared errors between the model output data and the output data, which is an actual measured value relating to the hull, is minimized during an identification value calculation period that includes at least the start time of the hull's change of course, and an update calculation unit that calculates coefficients of an update function, which is a regression equation relating to the hull's speed, based on an identification data sequence in which the calculated identification value and the evaluation quantity corresponding to the identification value are accumulated in time series, and calculates updated values ​​of the hull parameters using the update function to which the calculated coefficients are given.

The present invention relates to a method for determining insulation resistance (R1, R2) and discharge capacitance (C1, C2) for ground (PE) of an ungrounded DC power supply system, wherein a linear difference equation is implemented as a function of a measurement voltage (UG(k)) and a grid voltage (UN(k)) with a measurement voltage (UM(k)), and a grid parameter (θi) formed from a shunt (RM), insulation resistance (R1, R2) to be determined, and discharge capacitance (C1, C2) to be determined. Using N linear differential equations, a measurement value equation system is implemented with a sample sequence ((UG(k), UN(k), UM(k))) and a grid parameter (θi) for measurement times from k = 1, 2 to N with a sampling period T. Another method step comprises a step (S6) of calculating the estimated grid parameter (θi) as an approximate solution to the measurement value equation system, wherein the sum of squared errors between the grid parameter (I) and the estimated grid parameter (θi) is minimized, a step (S7) of minimizing the sum of squared errors through QR decomposition of the measurement-value matrix (Ψ), wherein the QR decomposition is recursively calculated, and a step (S8) of calculating each insulation resistance (R1, R2) and each discharge capacitance (C1, C2) from the estimated grid parameter (I) and repeating the method step with the estimated grid parameter (I), each calculated to include samples available at the current measurement time.

The invention discloses a low-voltage transformer area line impedance calculation and anomaly identification method based on information data of an electric energy meter and an electric terminal, and belongs to the technical field of analysis, operation and maintenance of a power system. The method comprises the following steps: carrying out anomaly detection and correction on power consumption data; based on the corrected power utilization data, low-voltage transformer area user-transformer relation identification is carried out by adopting DTW-CMeans clustering improved by a dream optimization algorithm; establishing a branch impedance model based on Kirchhoff's law according to the corrected power utilization data and the identified transformer area user-transformer relationship, and solving line impedance by using a cuckoo optimization algorithm and taking a root-mean-square error as a target; and based on historical impedance data, adopting a 3 sigma principle to carry out impedance anomaly identification. According to the method, the accuracy and the stability of user change relation identification are improved, the sensitivity and the estimation precision of impedance calculation to dynamic abnormal data are enhanced, and missed judgment and misjudgment in abnormal identification are effectively reduced through self-adaptive threshold setting.

The invention relates to a method for predicting the solubility of CO2 in ionic liquid based on multi-model fusion and interpretable machine learning, and the method comprises the steps: building a stacked integrated model based on a MindSpark deep learning framework, an integrated random forest and a gradient lifting regression algorithm, carrying out the automatic optimization of hyper-parameters through a Bayesian multi-objective optimization algorithm, and explaining a model prediction result through a SHAP value method. According to the method, the determination coefficient on a test set reaches 0.9823, the mean square error is 0.0010, the mean absolute error is 0.0210, and the root-mean-square error is 0.0323. The method solves the problems of tedious operation, time consumption and labor consumption in the process of measuring the CO2 solubility of the ionic liquid through experiments, overcomes the defects of insufficient precision, poor generalization ability and weak interpretability of a traditional prediction method, can rapidly and accurately predict the CO2 solubility of the ionic liquid, and provides theoretical guidance for molecular design.

Methods, systems, and devices for four-phase clocking techniques, and in particular for a hybrid clock cycle correction and quadrature error correction (DCC-QEC) circuit, are described. The DCC-QEC circuit comprises a DCC circuit configured to generate two-phase clock signals based on external clock signals. The DCC circuit is self-calibrating without receiving a reference signal.The DCC-QEC circuit further comprises one or more QEC circuits coupled to the DCC circuit, each of the QEC circuits comprising a pair of phase dividers configured to convert the two-phase clock signals into four-phase clock signals, and an offset matcher configured to match clock offsets associated with a subset of the four-phase clock signals based on QEC matching codes and in an open-loop manner, wherein data from multiple lanes of parallel data channels based on the offset-matched four-phase clock signals are serialized and sent to outside a device comprising the DCC-QEC circuit.

The invention belongs to the technical field of wireless communication and positioning, and particularly relates to a combined positioning and parameter calibration method based on user mobility and multipath assistance. According to the method, the problem of scene'rotation 'under the condition that base station and user antenna deviation exists is solved by utilizing displacement information of a known user, a nonlinear least square problem is constructed by utilizing measurement values obtained at two moments, the base station antenna deviation () and the target antenna deviation () are taken as known quantities to be substituted, and the target antenna deviation () is taken as the known quantities to be substituted. An initial value of a state quantity is solved by using a least square method, then estimation precision is further improved by using a Gauss-Newton iteration solution, then a cost function is put forward, a group of angles enabling the cost function to be minimum and corresponding position estimation are selected as initial state vectors of a subsequent Gauss-Newton iteration algorithm, and the estimation precision is further improved. And finally, constructing a complete state vector, and solving and optimizing an error caused by search by using Gaussian-Newton iteration. And finally, all state estimation results are obtained, and the root-mean-square error of each state is calculated.

The invention provides a wind speed prediction method based on a high-resolution wind speed model. The method comprises the following steps: acquiring to-be-predicted wind speed data; inputting to-be-predicted wind speed data into the wind speed prediction model to obtain a wind speed prediction result; the construction method of the wind speed prediction model comprises the following steps: carrying out bilinear interpolation processing on a wind speed data set to obtain an interpolated wind speed data set; according to a mask data extraction method, the interpolation wind speed data set is extracted, a data sample set is obtained, and the data sample set is divided into a training set and a test set; inputting the training set into the initial model to obtain a training result; calculating root-mean-square error loss according to the training result and the test set; according to the root-mean-square error loss, gradually correcting the initial parameters to be trained through an Adam optimization algorithm to obtain a training model; inputting the test set into the training model to obtain a wind speed simulation result; and evaluating the training model according to a wind speed simulation result to obtain a wind speed prediction model. The problem of low accuracy of wind speed prediction is solved.

The invention discloses a small sample surface roughness prediction method under a cross-working-condition background, and belongs to the technical field of machining. According to the method, cutting signals (force and vibration) and technological parameters (main shaft rotating speed, feeding speed and cutting depth) are collected in real time through a three-axis force sensor and a vibration sensor, and a multi-dimensional feature vector is constructed in combination with a multi-level feature extraction strategy (time domain statistics, frequency band energy spectrum and eight-layer wavelet packet time frequency entropy). A multi-head attention mechanism (MHA) is innovatively adopted to dynamically allocate feature weights, and the key feature characterization capability is improved. And a meta-learning domain generalization (MLDG) training framework is further proposed, a data set is divided by using target working condition information, a source domain data set is divided into a pre-training set and a fine tuning set, a real domain offset scene is simulated, and the robustness of small sample prediction of the model under cross-working conditions is improved. Experiments show that in a TB6 titanium alloy milling test, the surface roughness prediction determination coefficient R2 is 0.962, the mean absolute error MAE is 0.016 [mu] m, the root-mean-square error RMSE is 0.020 [mu] m, and the method has high prediction precision and reliability.

The disclosure deals with method and system for an over-the-air computation (OAC) approach for privacy-preserving localization in a sensor network. The disclosed approach relies on a voting-based distributed localization and the computation of the majority votes (MVs) with OAC. In this method, the anchor node (AN)'s votes encoding the potential location of a server node (SN) are mapped to a circularly-shifted chirp (CSC) signal and all AN simultaneously transmit their CSC signals. The aggregated signal is received at the SN and the MV votes are detected non-coherently without requiring an ideal time-synchronization or channel state information (CSI). We further disclose an iterative refinement procedure to increase the localization performance. The CSCs result in spectral-efficient and power-efficient transmission, while disclosed iterative refinements and repetitions decrease the gap between the root-mean-square error (RMSE) and the quantization bound.

The invention provides a blast furnace ironmaking multi-objective random optimization method, and relates to the technical field of intelligent optimization control in the blast furnace ironmaking process, and the method comprises the steps: firstly, building a nonlinear mapping relation among economic cost, CO2 emission, molten iron temperature, silicon content and decision variables based on operation data, taking the first two as performance indexes, and taking the latter two as constraint functions; obtaining a first joint probability density function of the performance indexes by adopting multi-dimensional kernel density estimation, and constructing a second joint probability density function representing an ideal target in combination with the correlation coefficient; taking a square error integral between the two as a performance optimization index, performing kernel density estimation on a constraint function, and introducing a chance constraint mechanism to form a probability type constraint condition; and finally, constructing a multi-objective optimization model, and performing iterative search on decision variables by using a differential evolution algorithm to realize multi-objective optimization of economical efficiency, low-carbon property and process feasibility.

Provided is a brushless direct current (BLDC) motor control method based on an improved proportional integral derivative (PID) controller, relating to the field of motor control. The method of this application comprises: establishing a squared proportional integral derivative (SPID) controller integrating a squared error term; determining a sign of a gain coefficient of the squared error term in the SPID controller based on an error between an actual speed and a target speed of a BLDC motor; determining whether to enable a derivative term of the SPID controller based on whether the actual speed has reached the target speed; using the error between the actual speed and the target speed as an input amount to the SPID controller, and outputting a current control amount; and performing speed control on the BLDC motor based on the current control amount until the error meets an accuracy requirement.