126 results about "Integration algorithm" patented technology

The invention discloses an underground engineering geological safety dynamic risk assessment method based on multi-source data fusion, which relates to the technical field of risk assessment, and comprises the following steps: collecting multi-source heterogeneous data related to underground engineering, extracting implicit information, modeling underground engineering geological safety risk factors into a risk network, and establishing a risk network model; calculating the comprehensive importance of the nodes based on a Stacking integration algorithm, and identifying key risk factors; acquiring characteristic parameters of key risk factors by using spatio-temporal characteristics of implicit information, introducing a random walk mechanism to acquire a dynamic accident prediction chain, and performing learning representation by using a graph attention network to acquire probability distribution of an accident evolution path; and assessing the vulnerability of the connection edge in the risk network, establishing a dynamic risk assessment model based on the node importance and the edge vulnerability, and obtaining a dynamic risk value corresponding to the accident according to the accident occurrence probability and the risk mitigation factor. According to the invention, intelligent identification, dynamic evaluation and accurate early warning of risk factors are realized, and the accuracy and real-time performance of risk identification and evaluation are improved.

The invention discloses a thermal barrier coating multi-dimensional terahertz nondestructive testing method and system based on an adaptive optimization mechanism, and the method comprises the steps: constructing a matrix sample library, and preparing thermal barrier coating samples with different thicknesses, porosities and uniformity; obtaining terahertz detection data of three dimensions of geometric dimension stability, microstructure integrity and material structure uniformity of the thermal barrier coating; establishing a multi-parameter correction model based on surface roughness, and performing feature extraction, fusion analysis and error compensation on the detection data by adopting a multi-modal integration algorithm; an adaptive optimization model is established through a deep learning algorithm, and continuous iterative optimization of the detection system is realized in combination with a feedback correction mechanism; the method is applied to an engine blade thermal barrier coating, an actual blade coating is detected, detection data are collected for system optimization, system parameters and models are updated, and system applicability and robustness are improved. The method can be widely applied to the field of nondestructive testing of coating structures such as aero-engine blade thermal barrier coatings.

The invention relates to the technical field of chemical safety management, and discloses an emergency equipment high-efficiency management system and method.The emergency equipment high-efficiency management method comprises the steps that multi-source heterogeneous data of emergency equipment in a chemical industrial park is collected and converted into a standardized data set through a heterogeneous data integration algorithm, and the standardized data set is stored in a database; the expert knowledge is converted into a domain knowledge graph by means of a hierarchical knowledge representation framework. Based on this, a multi-layer causal network model is constructed by using a causal discovery algorithm, analysis results and decision suggestions are extracted through multi-scale causal reasoning, and a dynamic causal knowledge base is generated through combination of a bidirectional calibration algorithm and newly added data; according to the method, the limitation of single analytic hierarchy process is broken through by fusing domain knowledge and multi-source data, explanation is provided for decision making, the method has adaptive evolution capability, and the efficiency and accuracy of emergency equipment management are greatly improved.

The invention relates to an AI server cooling water pump vibration quantity self-adaptive calibration test method and system, an intelligent test platform integrates a six-dimensional force sensor and a temperature compensation vibration exciter, the platform rigidity is automatically calibrated, a water pump-pipeline system transfer function is obtained, and a rotating speed-lift-modal frequency three-dimensional mathematical model is established; a three-axis MEMS accelerometer array is arranged at sensitive parts such as a water pump bearing seat and a motor shell, and vibration, current and pressure signals are synchronously collected; carrying out time-frequency domain signal processing by adopting variational mode decomposition in combination with self-adaptive S transformation, and extracting a 128-dimensional full-frequency domain feature vector containing a modulation side frequency band; effective values of vibration acceleration, speed and displacement are calculated through a frequency domain integration algorithm, and current harmonic interference is corrected through an electromagnetic vibration compensation model; and finally, a vibration health degree evaluation model containing 18 characteristic parameters is established based on a support vector machine, the test system comprises an intelligent test platform unit, a multi-source sensing unit, an edge calculation unit and a data management unit, and microsecond-level synchronous acquisition and GB-level data throughput are realized through a time sensitive network. And online incremental learning and automatic generation of a test report are supported. The method has the effect of improving the water pump vibration quantity test precision.

The invention discloses a GNSS (Global Navigation Satellite System) signal quality adaptive integrated navigation method, and relates to the technical field of vehicle navigation. IMU and GNSS positioning information and positioning states are read, integrated navigation is realized through an ESKF loose combination algorithm, a measurement covariance matrix is changed in real time according to the GNSS positioning state and a DOP factor, and finally, an error state value is estimated and fed back to an integrated navigation system. Position and speed information errors and attitude Euler angle errors under a navigation coordinate system are fed back to INS mechanical arrangement, angular acceleration zero offset errors of the IMU are fed back to the IMU, and updating of the state quantity is completed. A measurement covariance matrix is changed in real time according to a GNSS positioning state and a DOP factor by aiming at IMU and GNSS data through a loosely combined ESKF fusion algorithm, an error state is estimated and fed back to an integrated navigation system, and the positioning precision can be effectively improved.

The invention discloses a cement paste strength prediction method based on an integration algorithm. The method comprises the following steps: collecting original cement paste data; performing feature extraction on the original cement paste data to obtain to-be-predicted cement paste data; and sequentially inputting the to-be-predicted cement paste data into a trained convolutional neural network model, an XGBoost model and a random forest model to obtain the compressive strength of the cement paste test block with the standard size, so as to predict the cement paste strength according to the compressive strength of the cement paste test block with the standard size. According to the cement paste strength prediction method based on the integration algorithm, the strength of the cement paste is predicted by using three models, and respective advantages, the powerful capability of CNN in image and signal processing, the high efficiency of XGBoost in structured data processing and the robustness of a random forest in classification and regression problems can be utilized, so that the strength of the cement paste is predicted. Therefore, the prediction accuracy and robustness are improved.

The invention discloses a flexible finger friction contact collision dynamics simulation method based on an absolute node coordinate method, which comprises the following steps: simplifying a finger of a flexible manipulator into a connecting rod system, taking a tail end knuckle as a flexible beam for consideration, and carrying out unit division on the flexible beam system based on the absolute node coordinate method; a contact collision process is described by using an additional constraint method, meanwhile, switching of a viscous state and a sliding state is considered, a friction-containing flexible finger contact collision dynamic model is established, and a dynamic equation of a system is solved by adopting a double-circulation implicit integral algorithm; and obtaining a curve graph that the flexible finger tail end displacement changes along with time and the finger tail end contact force changes along with time. The invention provides a new frictional contact collision dynamic model for a flexible manipulator system, and provides more perfect data information and image information for researchers in the field.

The invention discloses a hierarchical database system tuning method and system based on a large language model, and relates to the technical field of database system tuning. The method comprises the steps that expert knowledge in the database field is processed based on a large language model, a cue word integration algorithm is used for converting the knowledge into structured knowledge capable of being used by a computer so as to optimize a search space, and a coarse granularity-fine granularity Bayesian optimization algorithm is used for determining optimal configuration, so that configuration optimization of a system level is completed; a parameter screening algorithm, a space pruning algorithm and a context Bayesian optimization algorithm are utilized, and an optimal execution plan is generated by selecting an appropriate optimization prompt for each query in the complete workload, so that configuration optimization of query levels is completed. According to the method, the search space is optimized by utilizing expert knowledge, the optimized space is explored by utilizing a Bayesian optimization algorithm, multi-level optimization is performed on the database system from a system configuration level and a query level, and the optimization efficiency and the system performance are improved.

The invention discloses a credit risk prediction method based on data and dynamic feature optimization and Bagging integration, and the method comprises the following steps: S1, obtaining an initial data set, carrying out the data preprocessing, balancing the data set through an ATGA algorithm, and reducing the data noise; s2, using LightGBM as an agent model, calculating importance scores of all features, gradually screening key features, adaptively adjusting feature screening standards according to data distribution through a dynamic threshold algorithm, gradually screening feature subsets, and inputting the screened features into a subsequent model; and S3, constructing a TabPFN model by using a Bagging integration algorithm, introducing a dynamic weight distribution mechanism, adaptively adjusting the weight of each base model according to sample features, and finally realizing credit risk prediction through the TabPFN model. According to the method, the credit risk prediction process is comprehensively optimized in three aspects of data set optimization, feature screening optimization and model performance optimization, so that indexes such as the accuracy rate and the rate of return are improved.

The invention relates to a single cell data integration method based on self-supervised learning, and the method comprises the steps: integrating single cell data from different researches through a single cell data integration algorithm scCorrector based on a research specific normalization layer, and mapping the single cell data to a public space; aligning a plurality of types of single cell data sets by adopting a single cell data integration algorithm scInterpreter which is based on knowledge regularization and is interpretable; integrating and annotating multi-species single cell data through a multi-species single cell data integration and annotation algorithm CAMEX based on a heterogeneous graph neural network; spatial transcriptional expressions are inferred from histological images using an algorithm stInfer based on a weighted K-nearest neighbor model. Compared with the prior art, the method has the advantage that adverse effects on downstream tasks caused by sequencing technology differences are relieved. In addition, gene expression is inferred through a dyed image, so that the research cost of a space transcriptome is greatly reduced, and convenience is provided for exploring a tissue microenvironment.

The invention discloses a retrieval enhancement generation method based on dynamic document block segment optimization, which comprises the following steps: S1, acquiring an input document, and preprocessing the input document; s2, the document is segmented into a plurality of paragraphs through a natural language processing technology, each paragraph comprises a plurality of smaller blocks, and an index is constructed according to the paragraphs and the blocks; s3, after a query request is received, related paragraphs and blocks are retrieved through indexes; s4, according to the requirement of the query request, the size of the blocks is dynamically adjusted, segments are constructed through the adjusted blocks, and each segment comprises a plurality of blocks; s5, fusing the block obtained by retrieval with the corresponding context information, and promoting the organic combination of the segment, the block content and the context information through a fusion algorithm; and S6, generating answers based on the fused segments through a language model, and performing inspection confirmation and content optimization on the generated answers.

The invention discloses a rock true triaxial aging damage mechanical model construction method, which comprises the following steps of: representing rock nonlinear rheological behaviors by adopting viscoplastic strain and aging damage, and reflecting a progressive equilibrium motion process in a rock creep process by utilizing an aging damage factor; a crack azimuth angle is considered, a damage variable is adopted to control and drive a yield surface, a rock true three-dimensional yield function is provided, and a rock true triaxial aging damage mechanical model is established; finally, an explicit nonlinear numerical integration algorithm is adopted, and creep process aging damage factor evolution is accurately calculated. The rock true triaxial aging damage mechanical model provided by the invention can accurately describe rock true triaxial instantaneous and creep mechanical response, provides a theoretical basis for long-term stability research of deep rock mass excavation engineering, and provides guarantee for safety and stability of deep rock engineering construction.

An online estimation method and system for state of charge (SOC) errors. The method comprises: acquiring an initial SOC error value at a power-on moment of a battery management system, and updating the initial SOC error value by means of an OCV table lookup algorithm when a correction condition is satisfied; calculating a real-time SOC error of a battery by means of an ampere-hour integration algorithm; detecting a real-time calculation process of the ampere-hour integration algorithm, and when a Rint model correction condition is triggered, correcting the real-time SOC error by means of a Rint model algorithm; and when the battery reaches a preset condition, zeroing out the SOC error and then performing updating by means of the ampere-hour integration algorithm until a power-off moment of the battery management system is reached, saving a corresponding real-time SOC error and outputting same. A theoretical basis for SOC error estimation is provided to realize online SOC error estimation; algorithms having higher precision are selected during estimation, so as to reduce SOC errors; and the calculation of a current SOC error can be calibrated, such that the method and system are adapted to scenarios involving different battery packs and different aging conditions.

The invention provides an artificial intelligence-based data analysis method, which comprises the following steps of: integrating and cleaning collected multi-source data, storing the data into a relational database, designing an entity-relationship-based data model to construct nodes and relationships, and storing the nodes and the relationships into the relational database; analyzing and calculating the knowledge graph through a community discovery algorithm, a graph neural network and quantitative analysis to capture complex features and potential abnormal conditions of the knowledge graph; and fitting three results of the community discovery algorithm, the graph neural network and the quantitative analysis through an integration algorithm, and generating final suspicious score judgment in a multi-party weighted voting mode. According to the method, suspected problem clues are extracted from thousands of or even hundreds of thousands of data, a possible clue direction is provided for case investigation, problems can be found more timely, supervision is more accurate and efficient, and case handling efficiency is improved.

The invention discloses a wavefront restoration method based on multi-path integration, and relates to the technical field of wavefront measurement and restoration, and the method comprises the following specific steps: data collection and initialization: carrying out the preliminary detection of wavefront through a Hartmann sensor microlens array, dividing the space where the wavefront is located into a plurality of rectangular small regions, according to the method, a multi-path calculation phase evaluation value set is introduced, all shortest paths from an initial point to each to-be-estimated point are found by using a shortest path search algorithm, it is ensured that multiple possibilities of wavefront propagation can be considered in the calculation process, and by introducing a path weight distribution mechanism, the estimation accuracy of the wavefront propagation is improved. According to the method, geometric characteristics, optical characteristics and correlation with surrounding paths of paths are comprehensively considered, each path is endowed with different weights, forward integration is carried out on discrete data results by adopting an innovative mixed integration algorithm, local detail loss or error amplification is effectively avoided, and thus a more accurate wavefront phase value is obtained.

The invention relates to an icing power transmission line galloping simulation method and simulation system based on a composite time integral. The method comprises the following steps: S1, establishing a nonlinear finite element model of an icing power transmission line; s2, carrying out time domain solution by adopting a composite time integration algorithm; calculating the solution of the moment in two sub-steps according to the solution of the moment, namely from the first sub-step to the second sub-step; the first sub-step adopts a trapezoidal rule for integration, and the second sub-step adopts a three-point backward difference format based on information of three moments; s3, coupling and solving a nonlinear aerodynamic load related to the speed; s4, obtaining the galloping response of the power transmission line; the problems that in a transmission line galloping simulation technology based on a traditional Newmark method, contradiction between calculation efficiency and numerical precision exists in the strong nonlinear coupling problem, accurate tracking of low-frequency galloping and efficient suppression of high-frequency noise are difficult to achieve at the same time under the large time step length, simulation calculation efficiency is low, and convergence is poor are solved.

The invention discloses an unmanned metamorphic vehicle parameter optimization method based on a multi-configuration grading and integration algorithm. The method comprises the following steps: 1, establishing an unmanned metamorphic vehicle model; 2, establishing evaluation indexes of the unmanned metamorphic vehicle in different states; 3, screening parameters by adopting an SOBOL sensitivity analysis method; 4, formulating a constraint domain of each parameter according to different-state performance and structural limitation; and 5, carrying out grading and integrated optimization on the screened parameters according to a sensitivity analysis result and a constraint domain. According to the method, the parameters which greatly influence the motion performance of the unmanned metamorphic vehicle can be screened and optimized, the optimization complexity is reduced, and the optimization time is shortened, so that the stability of the multi-configuration motion of the unmanned metamorphic vehicle can be improved, and the motion energy consumption is effectively reduced.

The invention discloses a method for predicting the lignin separation and extraction effect of a deep-eutectic solvent based on machine learning. The method comprises the following steps: constructing a sample database containing characteristic parameters such as hydrogen bond acidity, hydrogen bond alkalinity and polar surface area of the deep-eutectic solvent and a lignin removal rate; based on a solvent molecular structure, key characteristic variables influencing the lignin removal rate are extracted; a Stacking integration algorithm is adopted, Xgboost, RF and SVM are taken as base learners, CNN is taken as a meta learner, and a machine learning prediction model taking the removal rate as a target variable is constructed; performing hyper-parameter tuning by adopting five-fold cross validation, and constructing a machine learning prediction model; explaining the model by using an SHAP method, generating a feature importance ranking map and an SHAP value distribution map, and intuitively displaying the influence of key variables on the lignin extraction removal rate; finally, the lignin removal effects of different eutectic solvent systems are predicted through a machine learning model, and the solvent composition proportion is adjusted in combination with feature analysis so as to optimize the separation process.

The invention discloses a moon leap trajectory planning method, system and device and a medium. The method comprises the steps of obtaining target landing point data; based on the acquired data, establishing a moon surface leap device kinetic model by defining a specific coordinate system and a state variable; based on the kinetic model, discretizing the kinetic equation through a numerical integration algorithm, constructing and solving a leap trajectory convex optimization problem with optimal burnup, and obtaining an obstacle avoidance-free optimal trajectory; based on the obstacle avoidance-free optimal trajectory, performing obstacle avoidance iteration through an iteration obstacle increase strategy to obtain a static obstacle avoidance trajectory; and on the basis of the static obstacle avoidance trajectory, the flight trajectory is adjusted again by calculating the trajectory error band width and the maximum extension distance, and a dynamic safety enhanced trajectory is obtained. According to the method, safe, robust and efficient moon surface leap trajectory planning in a complex and uncertain environment is realized.

The invention relates to the technical field of data processing, and discloses an engineering construction whole process digital management method and system. The method comprises the following steps: collecting and standardizing engineering construction full-period multi-source data; constructing a stacked integrated prediction system by using the standardized data to obtain a three-dimensional index prediction value; allocating resources based on the predicted value in combination with the ergonomics risk index to obtain a configuration scheme; uncertainty factors are quantified for the configuration scheme, an optimal scheme set is calculated through a multi-target symbiotic algorithm, and a management strategy is formed. According to the method, high-precision prediction of the engineering three-dimensional indexes is realized through the stacking integration algorithm, ergonomics risk assessment and resource allocation optimization are combined, and the technical defect that human factor engineering is neglected in a traditional method is overcome.

The invention provides a photovoltaic array fault diagnosis method based on composite model soft voting integration. The photovoltaic array fault diagnosis method comprises the following steps: acquiring a plurality of P-V curve data when a photovoltaic array normally operates and fails; preprocessing the plurality of P-V curve data, and formulating fault type labels and performing division processing to obtain a training data set and a test data set; constructing a plurality of composite sub-models, integrating the composite sub-models based on a soft voting integration algorithm, determining various hyper-parameters of the integrated diagnosis model based on a global optimization algorithm, and completing the construction of the integrated diagnosis model; inputting the training data set into the integrated diagnosis model for training to obtain a trained integrated diagnosis model; inputting the test data set into the trained integrated diagnosis model to test and evaluate the performance of the integrated diagnosis model; and the photovoltaic array is diagnosed through the integrated diagnosis model, and the fault type of the photovoltaic array is determined.

The invention discloses an isogeometry-based composite material thin shell dynamics prediction method, and belongs to the field of composite material thin shell dynamics behavior prediction. The implementation method comprises the steps that geometric modeling is conducted on the composite thin shell based on an isogeometric analysis method, unification of geometric modeling and numerical analysis is achieved through a non-uniform B spline (NURBS), and a composite thin shell geometric model is obtained; calculating a tensile stiffness matrix, a tensile-bending coupling stiffness matrix and a bending stiffness matrix of the composite material thin shell through the composite material parameters; establishing a kinetic model of the composite material thin shell unit through continuous medium mechanics, assembling the kinetic model to obtain an overall kinetic model of the composite material thin shell, and establishing a kinetic equation of the composite material thin shell based on the overall kinetic model of the composite material thin shell and the stiffness matrix; a kinetic equation is solved based on a generalized alpha implicit time integration algorithm, the kinetic behavior of the composite thin shell under the action of an external force F is predicted, and the stability and precision of thin shell kinetic behavior prediction are improved.

The invention relates to the technical field of indoor layout design, and discloses an indoor layout design system based on user behavior data analysis. The system comprises a user behavior data acquisition module which captures and transmits a user movement track coordinate sequence, residence time distribution data and equipment interaction event records in real time through a multi-sensor network; the layout analysis module constructs a dynamic layout model by using a decision tree integration algorithm, and outputs theoretical layout parameters including a space occupancy rate and functional region division; the layout difference detection module performs multi-dimensional comparison of space overlapping degree, path conflict index and regional utilization efficiency difference on theoretical and actual layout parameters to generate a difference index vector; the problem region positioning module identifies a layout abnormal region through a spectral clustering algorithm based on the difference index vector and predefined spatial topology information; and the adaptive design strategy generation module automatically generates a layout adjustment strategy instruction according to the abnormal region position and the attribute data.

The invention discloses a three-dimensional design model conformity analysis method and system based on a large language model, and the method comprises the steps: obtaining the related data and design specification constraint conditions of a three-dimensional design model, and building a data interaction channel through a collaborative review platform; calling a semantic geometry collaborative verification algorithm to process the data and extracting feature parameters; generating a rule adaptation weight matrix through a space rule learning adaptation model; constructing a mapping relation between the feature vector and the design specification through a CAD point cloud feature coding mapping model; carrying out multi-dimensional conformity verification on the basis of the result; and outputting the comprehensive analysis data set. The system realizes deep cooperative processing of geometric and semantic data of a three-dimensional design model through multi-unit cooperative linkage and integration of an algorithm and a platform function, improves the accuracy of rule adaptation and feature mapping, comprehensively covers a conformity verification dimension, and meets a high-precision analysis requirement in a complex scene.

The invention provides a TCN-GRU-Adaboost-based load prediction method, which is applied to the field of load prediction of an integrated energy system. Firstly, a time sequence signal is simply decomposed and reconstructed; secondly, in combination with a time sequence convolutional network (TCN), a gating cycle unit (GRU) and an attention mechanism (Attention Mode), establishing a hybrid neural network; then combining an adaptive enhancement integration algorithm (Adaboost) with the hybrid neural network to form a load prediction model; and evaluating power, gas, heat and cold load data of a certain integrated energy system by using the constructed load prediction model. Compared with the prior art, the method is higher in precision, and better in robustness and generalization of prediction of different types of loads of the integrated energy system.

The invention discloses a bucket dividing regulation and control method and device for a financial data warehouse and electronic equipment, and relates to the technical field of artificial intelligence or other related fields, and the method comprises the steps: collecting a time sequence performance set of the financial data warehouse through a distributed collection system; multi-level feature extraction is carried out on the time sequence performance set by adopting a sliding window mechanism to obtain a time sequence feature set, and the categories of time sequence features comprise instantaneous fluctuation features, periodic trend features and steady state baseline features; performing association analysis on the time sequence feature set based on a decision tree integration algorithm by taking an expected query performance requirement as a target to obtain a bucket-dividing adjustment strategy; and bucket index reconstruction is carried out on the financial data warehouse based on the bucket division adjustment strategy, and the financial data warehouse after index updating is obtained. Through the method and the device, the technical problem that the system performance is degraded due to the fact that a static bucket dividing mechanism of a real-time data warehouse cannot dynamically adapt to space-time imbalance of data loads in a financial service scene in the related technology is solved.

The invention relates to the technical field of neural signal processing, in particular to a behavior neural signal optimal weighted integration algorithm applied to language brain diseases, which comprises the following steps of: 1, acquiring a basic activity mode of a brain of a user through electroencephalogram data, and analyzing characteristics of low-frequency oscillation and a functional connection network; 2, in task state electroencephalogram data collection, a user completes a language processing task; step 3, identifying two sets of electroencephalogram indexes closely related to language brain diseases through feature extraction and screening; 4, through an iterative optimization method, different weights are given to different indexes, and the reasonability of weight distribution is verified through a machine learning technology; and 5, generating a comprehensive language brain health score for clinical evaluation. According to the method, multi-dimensional data are integrated into index weight distribution of neural signals, and brain multi-dimensional electrical data and behavioral data are combined, so that a set of efficient language brain function evaluation system is constructed.

The invention discloses a customer service event dynamic scheduling method based on time sequence modeling, and the method comprises the following steps: S1, collecting heterogeneous original data, and constructing a service event sequence; s2, inputting the service event sequence into an improved Transform model, and outputting an event semantic feature vector; s3, constructing a multi-scale time window, and generating a time feature vector array; s4, splicing the event semantic feature vector and the time feature vector to generate a joint feature vector; s5, inputting the joint feature vector into an improved decision tree integration algorithm, generating a scheduling priority score value, and recording a decision path and a corresponding feature node index; s6, executing a service event queue operation, and generating a service event scheduling list; and S7, acquiring service event execution feedback information, and updating parameters. According to the method, time sequence modeling, feature scoring and feedback optimization mechanisms are fused, and intelligent recognition, efficient sorting, interpretable scheduling and dynamic evolution of customer service tasks are achieved.

The invention provides a training method for aero-engine modeling driven by mechanism and data fusion, and relates to the technical field of aero-engine modeling simulation, and the method comprises the steps: decomposing a simulation model of an aero-engine into a plurality of assemblies containing to-be-identified unknown parameters, and employing an assembly integration algorithm to associate the plurality of assemblies, an end-to-end differentiable complete simulation model is formed; based on hierarchical division of the simulation model, performing parameter pre-estimation by adopting a top-down layer-by-layer progressive strategy to obtain an initial estimation value of the unknown parameter; the strategy comprises the following steps: firstly, estimating unknown parameters of components in a highest level according to actually measured output data, and generating a training target of a next lower level based on a calculation result of an estimated level so as to complete estimation of unknown parameters of the lower level in sequence; and performing gradient back propagation in the complete simulation model by taking the initial estimation value as a starting point to optimize and update unknown parameters.

The present invention discloses a method for identifying spatial micro-motion targets based on the Stacking integration algorithm, which mainly solves the problems of single feature extraction in existing spatial micro-motion target identification methods and the inability of traditional single classifiers to fully exploit the classification potential of features. The implementation scheme is as follows: 1) Use the multi-transform domain feature extraction method to extract the target time domain, frequency domain, and time-frequency domain features, and generate a training sample set and a test sample set; 2) Construct a Stacking integration classifier model in which four primary classifiers are connected in parallel and then cascaded with a secondary classifier; 3) Use the training sample set and the cross-validation method to train the Stacking integration classifier; 4) Input the test sample set into the trained Stacking integration classifier to obtain the classification result. The present invention can fully characterize the target characteristics, exploit the classification potential of features, improve the recognition rate of spatial micro-motion targets, and can be used for ballistic target recognition.