58results about How to "Achieve forecast" patented technology

The invention belongs to the technical field of multi-rotor unmanned aerial vehicle control, and particularly relates to a multi-rotor unmanned aerial vehicle self-adaptive feedforward control method based on a mesoscopic wind field model. Comprising the following steps: firstly, based on a relaxation time model of a lattice Boltzmann method, establishing a mapping relation between a Knudsen number and turbulence intensity and thermal noise intensity of a macroscopic wind field, and realizing parameterized representation of a mesoscopic wind field; then, constructing a mesoscopic wind field model by synthesizing a real-time wind field containing average wind, turbulent flow, gust and thermal noise components; then, calculating the wind resistance according to the predicted wind speed, designing a self-adaptive gain mechanism fusing a real-time tracking error, an error wind direction included angle and a historical error trend, and generating a dynamic feedforward control quantity; and finally, combining the feedforward control quantity with a feedback control quantity based on gravity compensation to form a comprehensive wind resistance control law. Starting from the mesoscale, the disturbance suppression capability and trajectory tracking precision of the unmanned aerial vehicle in complex environments such as strong wind and turbulent flow are enhanced.

The application discloses an autonomous action prediction method based on a sequential space-time domain brain and muscle electric signal, first, a space-time graph network matrix is constructed, space-time data mapping is performed on the collected brain and muscle electric signal of a patient, and the signal is converted into a graph network matrix corresponding to a plurality of continuous time frames in a data window division manner; then, a graph network matrix of a future signal is predicted, the graph matrix data of the plurality of continuous time frames are sent into a graph convolutional neural network for analysis, and space-time data prediction is performed on the graph matrix corresponding to the future time frames; finally, target action recognition is performed, the predicted plurality of time frame graph matrices are subjected to pattern recognition through the graph convolutional neural network containing a plurality of full connection layers, the action corresponding to the signal is analyzed, and the corresponding action of a control auxiliary device is output in advance. The relatively high explainability of the algorithm can help popularize and apply the algorithm in the rehabilitation medical field, thereby helping more stroke patients.

The application provides a network security vulnerability automatic management method based on network security intelligence, belongs to the field of network security vulnerability management, and is used for solving the problems of low automation, large risk assessment deviation and poor collaboration in the related art. The method comprises network security intelligence collection, screening, storage and analysis, attack path prediction, key asset identification, hazard scene enhancement and vulnerability management. Through technologies such as dynamic weight, high-order probability graph model and multi-agent game, accurate intelligence processing, dynamic risk quantification, attack forward-looking prediction and cross-enterprise collaborative disposal are realized, the vulnerability management efficiency and accuracy are improved, and the network security of the ECUs is ensured.

The application discloses a kind of based on sequence stratigraphy and seismic sedimentology's braided river channel reservoir fine characterization method, comprising the following steps: obtaining target area core, well logging and high-resolution three-dimensional seismic data;By analyzing the sedimentology mark and logging facies mark in the target area, to determine the sedimentary facies type and sedimentary microfacies division;Using lithology and well logging curve, using the theory of sedimentology, high-resolution sequence stratigraphy identifies sequence interface type and feature, and carries out high-resolution sequence stratigraphic sequence division;Division reservoir architecture interface;According to the internal structure of sedimentary system, sedimentary facies distribution, focus on high-frequency sequence and sedimentary body scale, generate the braided river channel reservoir fine characterization result of target area architecture unit.The present application effectively solves the deficiencies of prior art in reservoir fine description, boundary identification and parameter inversion, improves the efficiency and success rate of oil and gas reservoir development.

The invention discloses a dynamic motion primitive coding method and system based on Lie group Lie algebra, and the method comprises the steps: uniformly representing the tail end pose and contact force spinor of a robot in a Lie group form, and constructing a unified mathematical framework; the method comprises the following steps: acquiring a pose and a force spinor sequence through a teaching phase, and converting the force spinor sequence into a virtual pose sequence through scaling mapping; establishing a dynamic motion primitive (DMPs) model based on the Lie group Lie algebra theory, and learning a model weight matrix and a primary function parameter by adopting local weighted regression; and according to the initial and target poses of the new task, generating a smooth trajectory through phase synchronization control and Euler integration, and outputting the smooth trajectory to a robot controller. According to the method, the problems of posture singularity and motion distortion caused by splitting of a posture channel in a traditional DMPs method are solved, the learning fidelity and reproduction stability of the six-degree-of-freedom posture track of the robot are improved, the smoothness and coordination of the whole motion process are guaranteed, and the debugging complexity of a force control system is reduced.

The invention relates to a regional ocean sound propagation field millisecond-level prediction method based on a deep neural network, and the method comprises the steps: constructing two factors which have the greatest influence on a sound propagation field for a target region: a sound velocity profile and a terrain; an environment information data set of an area sound propagation field is constructed by constructing an annual sound velocity profile of a target area node and building a large number of two-dimensional terrain models by extending around the node by 360 degrees, and the sound propagation fields of the area under different environment conditions are calculated by using a traditional sound field calculation model to serve as a training set; feature simplification parameters of the sound velocity profile and the terrain in each environment sample are extracted to serve as guide information of a corresponding sound propagation field, and a deep neural network is built to learn distribution rules of the sound propagation fields under different environment conditions in the area; and finally, the trained model can quickly predict the corresponding sound propagation field in the region according to different environmental condition characteristic batches.

The invention relates to the technical field of solid lasers, and particularly provides a narrow-linewidth tunable dual-wavelength laser output system based on data driving, which comprises a pumping source, a resonant cavity, a data acquisition unit and a data receiving and driving device, a single gain medium, two F-P etalon and an F-P angle regulator are arranged in the resonant cavity, the inclination angles of the two F-P etalon are adjusted through an F-P angle adjuster so as to perform longitudinal mode selection and wavelength tuning on the output laser of the resonant cavity; a neural network screening model is arranged in the data receiving and driving device, the neural network screening model is used for learning the mapping relation between intracavity parameters and output laser optical parameters, and dual-wavelength simultaneous locking and line width prediction are achieved through dual-etalon collaborative mode selection and data driving of the neural network screening model. According to the invention, the complexity of the system is greatly simplified, and dual-wavelength, narrow-linewidth, high-stability, repeatable and automatic laser output regulation and control which cannot be achieved by a traditional architecture are realized functionally.

The application provides a method and system for predicting main steam parameters before steam warming of a combined cycle unit, the method comprising the following steps: obtaining historical operation data of a target power generating unit, the historical operation data comprising a plurality of observable input parameters and output parameters closely related to a main steam system; constructing an NARX neural network model, training the neural network model based on the historical operation data, and obtaining a trained main steam parameter prediction model; deploying the main steam parameter prediction model on a general computing device; obtaining initial state parameters under a current working condition before starting the power generating unit, and inputting the initial state parameters into the main steam parameter prediction model; and outputting a predicted trajectory of main steam key parameters before the high-pressure bypass system is put into operation from the main steam parameter prediction model.

The present application relates to a kind of marine diffusible matter concentration field prediction method, system and equipment, belong to marine environment monitoring technical field.It includes: respectively deploying acoustic sensor node cluster to each sub-region of diffusible matter distribution, and constructs multiple NG-RC module;The acoustic sensor node in each acoustic sensor node cluster is divided into prediction and observation sensor node, and corresponding original observation sequence is obtained;Input vector is obtained according to original observation sequence, the vector is input into NG-RC module, and the local concentration prediction result of each sub-region is obtained;Cross-regional association information fusion is carried out to each NG-RC module, and according to local concentration prediction result, the global prediction result of each sub-region is obtained;According to the spatial position of each sub-region, the global prediction result of each sub-region is spliced, and the prediction concentration value of corresponding time whole region is obtained.The present application improves timing prediction ability and improves prediction precision, and enhances the adaptability to complex environment.

This application provides a method and apparatus for diagnosing the metal wear condition of transformer power components. The diagnostic method includes: constructing a distribution function of ferromagnetic and non-ferromagnetic particles in the transformer; acquiring data collected in a first period; inputting the collected data into the distribution function of ferromagnetic and non-ferromagnetic particles to calculate ferromagnetic and non-ferromagnetic particle data; adding the ferromagnetic and non-ferromagnetic particle data to obtain the metal particle data for the first period; acquiring the metal particle data for the second period; calculating the median particle size and weight growth rate of the metal particles based on the metal particle data from the two periods; and determining the metal wear condition of the transformer power components based on the median particle size and weight growth rate. By collecting and statistically analyzing metal particles in the oil of a faulty transformer, a statistical distribution function of the metal particles is obtained, and combined with actual monitoring data, prediction of small-diameter metal particle data that cannot be directly measured is achieved.

The invention provides a tool wear state monitoring method and device based on semi-supervised learning and a medium. The method comprises the steps that a multi-mode tool wear data set and a semi-supervised teacher-student learning model are constructed; performing forward reasoning on the multi-modal data without the labels according to a teacher model to obtain a plurality of pseudo labels; co-training the student models; generating a confrontation network and a dynamic data enhancement strategy according to a preset condition, and processing the semi-supervised teacher-student learning model; according to an integrated learning strategy of the model snapshot, migrating a pre-trained model weight to an image branch of the semi-supervised teacher-student learning model so as to optimize the semi-supervised teacher-student learning model; and inputting the multi-modal data of the to-be-detected tool in the machining process into the semi-supervised teacher-student learning model to obtain a wear state prediction result corresponding to the to-be-detected tool so as to realize high-precision and strong-generalization real-time monitoring and prediction of the wear state of the tool by fusing the multi-modal sensing information and the visual information.

The invention specifically discloses a self-adaptive multi-stage residual life prediction method and device for a bearing, and belongs to the technical field of life prediction. The method comprises the following steps: performing time-frequency domain analysis processing on vibration signal data of a full life cycle of a bearing, and dividing the data of the full life cycle into a healthy stage, a slow degradation stage and an accelerated degradation stage; putting the obtained multi-stage data into a dense connection network for supervised learning, and then performing adaptive stage division on an unlabeled data set of a target domain by using a prediction model obtained by training; the method comprises the following steps of: training a basic model by taking health stage data as a source domain, and realizing feature alignment to a target domain through a multi-kernel maximum mean difference (MK-MMD) method; a layered parameter freezing and fine tuning strategy is adopted, and data in the slow degradation stage and the accelerated degradation stage are gradually integrated for model optimization. According to the method, time migration of the model between different degradation stages and space migration of the model to the target domain are considered, and prediction of the whole life cycle RUL under the complex working condition is achieved.

The invention provides a tidal current sediment motion prediction method and system based on deep learning, and relates to the technical field of deep learning, and the method comprises the steps: calling a pre-trained tidal current sediment prediction model to carry out the preliminary prediction of tidal current sediment associated data containing tidal current power associated information and sediment characteristic associated information, and generating an initial prediction result; acquiring actual observation data and an initial prediction result, performing deviation analysis to obtain deviation characteristics, inputting the deviation characteristics into a parameter adjustment module to dynamically adjust model network parameters to obtain an optimization model, calling the optimization model to perform prediction again to generate an intermediate prediction result, and finally performing iterative optimization in combination with the deviation characteristics to obtain an optimal prediction result. According to the method, the prediction deviation is continuously corrected through an iteration updating mechanism, a final accurate tidal current sediment movement prediction result is obtained, dynamic and accurate prediction of tidal current sediment movement is achieved, interaction and real-time changes among complex factors are fully considered, and the prediction reliability is greatly improved.

The invention discloses a supply chain ESG risk dynamic assessment method and system based on multi-source data fusion, and the method comprises the following steps: 1, collecting multi-source heterogeneous data at each node of a supply chain, processing the data, and generating a node data set containing ESG risk features; 2, based on the node data set generated in the step 1, in combination with attribute parameters of supply chain nodes and external dynamic information, adaptively generating a dynamic evaluation benchmark; and step 3, on the basis of the node data set generated in the step 1 and the dynamic assessment benchmark generated in the step 2, constructing a risk conduction model to carry out real-time risk assessment, and outputting an assessment result containing a risk conduction path. And multi-source implicit information is fused by combining a space and semantic association technology, so that the objectivity and comprehensiveness of an evaluation basis are ensured.

The present application relates to the technical field of tree species optimization screening, and discloses a tree species multi-element mixed carbon sequestration optimization screening method based on artificial intelligence, which comprises the following steps: constructing the single growth characteristics and mixed growth characteristics of tree species according to the pretreated multi-source growth data; predicting the carbon sequestration amount of different tree species combinations in a given growth environment by using a carbon sequestration amount prediction model; calculating the joint growth adaptability and functional complementarity of different tree species combinations, comprehensively scoring the tree species combinations, screening a candidate tree species combination set based on the comprehensive score, constructing a multi-objective optimization function of the candidate tree species combinations in the candidate tree species combination set and solving the function to screen the candidate tree species combination with the highest carbon sequestration efficiency. Through multi-source growth data preprocessing and feature construction, tree species combination carbon sequestration amount prediction, joint scoring and screening and multi-objective optimization, the present application realizes the quantitative evaluation of the carbon sequestration potential of different tree species combinations and the selection of the optimal tree species combination configuration.

The present application belongs to the technical field of reliability evaluation, and in particular to a multi-section beam structure reliability evaluation method, device, equipment and medium. For the reliability test data of the multi-section beam structure under the qualitative and quantitative mixed factors, a UD learning function is used to gradually increase the high-quality training samples by comprehensively predicting the uncertainty of the structure performance and the qualitative and quantitative mixed factor distance factor, the Kriging model constructed based on the initial training set is iteratively updated until the set prediction relative error termination criterion is reached, and finally the prediction of the failure probability, a reliability index, is realized. The present application can realize efficient and accurate beam structure reliability evaluation with as few training samples as possible.

According to the shield tail sealing simulation test device and test method, a shield tail structure is truly restored, and compared with an existing reduced scale model test device, the shield tail sealing simulation test device and test method are more suitable for actual engineering. The shield tail brush spacing, the number of sealing cavities and the shield tail gap can be changed according to parameters required by different projects by flexibly adjusting the structure of the device, shield tail structures of different projects are subjected to simulation reduction, engineering adaptability inspection of shield tail sealing grease is carried out, a basis is provided for model selection of shield tail sealing grease in shield tunnel construction, and the construction efficiency is improved. The problem that in current tunnel construction, type selection of shield tail sealing grease mainly depends on engineering experience is solved.

This invention relates to a method for determining soil compaction degree using balers based on real-time machine parameters. Specifically, it relates to a method for determining soil compaction degree in the agricultural field. The purpose of this invention is to address the problems of existing soil compaction degree determination methods, which rely on laboratory testing, consider only a single parameter, have low accuracy, and poor real-time performance. The process is as follows: Acquire the real-time driving speed of the tractor, the tractor's own weight, the tractor's load weight, the tractor's tire pressure, the number of tractor trips, the mass of straw collected per unit time, and the baler's straw compaction density; obtain standardized real-time driving speed of the tractor, standardized tractor's own weight, standardized tractor's load weight, standardized tractor's tire pressure, standardized number of tractor trips, standardized mass of straw collected per unit time, and standardized baler's straw compaction density; calculate the soil compaction degree using a matrix method; and classify the soil compaction degree into grades.

The invention relates to the technical field of balance testing of structural components, in particular to a digitization-based bearing stability evaluation method for a spandrel girder structure, which specifically comprises the following steps of: constructing a virtual reality feedback model; performing internal micro-damage analysis caused by physical continuity attenuation, extracting an internal recessive stripping index obtained through the internal micro-damage analysis, and importing balancing fluctuation data of the bilaterally symmetrical support obtained by combining real-time weight measurement counterforce into oblique displacement offset plug evaluation analysis to perform bearing stability proofreading; evaluating the deterioration condition of the connection rigidity of the interface structure; and according to the oblique displacement offset plug index and the structural interface stability evaluation index, carrying out joint anomaly judgment to obtain a comprehensive collapse pre-estimation index and carrying out risk alarm. According to the method, the problem that in the prior art, before visible damage or remarkable deformation occurs to the bearing beam with the component stress specificity, an effective technical means capable of early warning sudden instability (such as oblique section shear damage) caused by internal damage accumulation of the bearing beam in advance is lacked is solved.

The invention discloses a data modeling-based low-light night vision optical performance test system, and relates to the technical field of photoelectric instrument performance detection and intelligent diagnosis, and the system comprises a target plate module which is used for providing an observation target with a division pattern for a detected low-light night vision instrument; and the image acquisition and processing module is used for acquiring an image output after the detected low-light night vision instrument observes the target plate module, and processing the image to extract image feature data. According to the data modeling-based low-light night vision optical performance test system, through automatic image acquisition and processing and in combination with a multi-algorithm parallel computing and cross verification process, a traditional method depending on visual judgment of an operator is replaced, the influence of subjective factors and environmental fluctuation on a detection result is effectively reduced, and the detection accuracy is improved. And the consistency and objectivity of the detection result are improved. The system can automatically complete the whole process from image acquisition, feature extraction to parameter verification, and the efficiency and reliability of detection operation are improved.

The invention discloses a method for predicting the hydrogen content in steel in a vacuum refining process. The method comprises the following steps: collecting initial working condition parameters of vacuum degassing; calculating the weight and the circulating flow of molten steel in a steel ladle-vacuum chamber double-phase region; solving real-time dehydrogenation rates and dehydrogenation cumulants of three dehydrogenation paths including the interior of the molten steel, the free surface of the molten steel and the surface of the argon bubble in the unit time step length; based on a mass conservation equation, establishing a transmission equation of the hydrogen concentration in the steel ladle-vacuum chamber two-phase region steel liquid circulation process, and iteratively solving the hydrogen concentration of the steel ladle and the vacuum chamber steel liquid in unit time by utilizing a fourth-order Runge-Kutta method; and outputting the hydrogen content of the molten steel in the steel ladle and the predicted value of the real-time hydrogen content change by iterating the time step length. The method is suitable for SSRF, RH and other vacuum refining devices, the absolute error between the prediction result and the detection value of the online hydrogen meter does not exceed 0.5 ppm, and accurate data support can be provided for process optimization.

The invention relates to the technical field of array antennas, in particular to a high-efficiency low-sidelobe array antenna optimization method, system and device and a medium. The method comprises the following steps of: obtaining directional diagram simulation data of a typical array element by using a full-wave electromagnetic simulation method, and training a residual neural network by using physical parameters of the typical array element and the directional diagram simulation data to obtain a typical array element residual neural network model; the method comprises the following steps of: training other general array elements by using simulation data and a fine tuning technology to obtain a general array element residual neural network model, then obtaining directional diagram prediction data by using the trained neural network model, optimizing physical parameters and excitation amplitude by using a BES algorithm, and finally putting optimized variables into a full-wave electromagnetic simulation method for verification; according to the method, the high-precision residual neural network model is trained through a small amount of simulation data and transfer learning, the time is greatly saved, meanwhile, when the BES algorithm is used for optimization, a simulation method does not need to be repeatedly called, only the network model needs to be called, and the optimization efficiency is very high.

The invention discloses a physical informed machine learning high-entropy alloy phase prediction system and method based on semi-empirical parameters, and belongs to the technical field of crossing of high-entropy alloy material design and machine learning. The system takes semi-empirical parameter physical constraint as a core, integrates the physical interpretability of an empirical parameter method and the data driving advantages of machine learning through a three-stage cooperation mechanism of'feature multiplexing-independent prediction-Bayesian fusion ', and solves the problems of narrow phase coverage, low multi-phase prediction precision, 'black box' defect and the like of a traditional method. The system can predict more than 10 high-entropy alloy phase types and multi-phase coexistence systems, the single-phase prediction accuracy rate in 856 groups of multi-component high-entropy alloy test sets reaches 100%, the multi-phase coexistence system prediction accuracy rate reaches 77%, phase formation physical mechanism explanation can be output, B2 phase exclusive accurate criteria are provided, and high-entropy alloy design is promoted to be transformed from a trial and error method to an accurate prediction method.

The application discloses a vehicle early warning method, device, equipment and medium. The method comprises the following steps: when a vehicle to be warned enters a specified area, acquiring vehicle position information of the vehicle to be warned, vehicle state information of surrounding vehicles, and preset area level information of the specified area; determining target vehicles satisfying early warning conditions and corresponding early warning event intensity information according to the preset area level information, the vehicle position information and the vehicle state information; and issuing the early warning event intensity information to the vehicle to be warned. The area level of the vehicle to be warned and the surrounding vehicles satisfying the early warning conditions is divided according to the preset area level information, and then the early warning event intensity information is determined according to the area level at the division position, so that the vehicle to be warned is warned with different early warning intensities. Real-time perception and prediction of abnormal driving behaviors are realized, and differentiated information prompting and early warning are realized. The safety risk near an expressway exit ramp is reduced, and the traffic efficiency is improved.

The invention discloses a monitoring, regulating and controlling system and method for additive-equal-subtractive composite manufacturing, and relates to the technical field of additive manufacturing. The defect monitoring module is used for predicting the defect generation condition of the future formed part; the stress deformation prediction module is used for predicting a future stress field evolution sequence and / or deformation field evolution sequence; the process coordinated regulation and control module is used for generating additive manufacturing parameters and / or hierarchical control instructions of the additive-equal-subtractive manufacturing module in the next period according to the defect and stress / deformation field prediction result; the additive-equal-subtractive manufacturing module is used for executing additive or equal or subtractive manufacturing operation; through global data acquisition, multi-process defect prediction, stress deformation prediction and a hierarchical coordinated regulation closed-loop architecture driven by a prediction result, prospective accurate management and control of defects and stress deformation in the additive and subtractive composite manufacturing process are realized, and the forming quality and the yield of complex components in additive and subtractive composite manufacturing are improved.

The present application relates to the technical field of electric digital data processing, and particularly relates to a tea sample processing method. The method comprises the following steps: acquiring a meteorological parameter list R of tea samples, R=(r1, r2, …, r i ,…,r u ), r i is a record of the meteorological parameter of the i-th tea sample; acquiring a standard priority sequence Z of the tea samples; traversing R, and adding r j i,m to a preset j-th original sequence H m,j of the m-th meteorological parameter; acquiring a Pearson correlation coefficient original list P; traversing P, and if p m,j ≥p0, then adding the key information (m, j) corresponding to p m,j to a preset target parameter sequence G; acquiring a target weight sequence A; acquiring data E of a target tea; traversing E, and according to e n , acquiring the grade b n of the target tea; and acquiring the priority z' of the target tea. The present application can automatically judge the priority of tea.

The application provides an austenitic stainless steel casting in-situ micro-alloying modification method based on multi-field coupling and intelligent feedback, comprising the following steps: S1: mold presetting and machine learning model deployment; S2: master alloy smelting and pouring; S3: injecting a micro-alloying agent into the mold and carrying out multi-field coupling treatment on the molten steel; S4: predicting the microstructure evolution of the casting through a machine learning model, and adjusting the micro-alloying agent injection parameters and the multi-field coupling parameters in real time; S5: carrying out solid solution treatment and aging treatment on the solidified casting to obtain a final product. The application synchronously carries out the micro-alloying process and the solidification process, the multi-field coupling effect of the pulsed electromagnetic field and the ultrasonic field, and introduces a machine learning driven intelligent feedback system, so that not only the processing steps are provided, but also an intelligent casting system which can be self-adaptive and self-optimized is constructed, and a new technical path is provided for the performance customization production of high-end castings.

This invention provides a photovoltaic (PV) operation status management and control method and system, comprising: constructing and training a PV ecological monitoring model including a physical coupling field submodule and a PV ecological agent submodule; performing risk assessment on a real-time PV operation dataset based on the PV ecological monitoring model to generate risk assessment results; performing strategy exploration and counterfactual reasoning based on the risk assessment results to generate corresponding optimal control strategies; executing the optimal control strategies; and performing feedback optimization based on the feedback data corresponding to the optimal control strategies, thereby achieving efficient and accurate management and control of PV operation status.

The application relates to a business data processing method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring a historical business data sequence arranged in a time sequence by a plurality of historical business data; determining a historical period to which the historical business data belongs, acquiring reference business data sequence of a reference period corresponding to the historical period, and obtaining first comparison business data based on the reference business data sequence; acquiring a first business data difference between the historical business data and the first comparison business data; acquiring a reference data difference, determining an abnormal scene influence value corresponding to the historical business data based on the first business data difference and the reference data difference; arranging the abnormal scene influence value corresponding to the historical business data in a time sequence to obtain a scene influence value sequence; and determining target business data based on the historical business data sequence and the scene influence value sequence. The method can improve the accuracy of predicted business data.

This invention belongs to the field of pedestrian trajectory prediction technology, and particularly relates to a method for predicting pedestrian trajectories at unsignalized right-turn intersections based on game theory. The method includes the following steps: S1, acquiring historical data on pedestrians and vehicles at unsignalized right-turn intersections; S2, analyzing the human-vehicle game factors at unsignalized right-turn intersections and constructing a corresponding human-vehicle game model; S3, inserting the human-vehicle game model into a pre-set S-GAN model to obtain an SDG-GAN model for predicting pedestrian trajectories; S4, training the SDG-GAN model using the historical data acquired in S1; S5, using the trained SDG-GAN model to predict pedestrian trajectories at unsignalized right-turn intersections in real time. This invention ensures the accuracy of pedestrian trajectory prediction at unsignalized right-turn intersections and the effectiveness of assisted driving decisions at such intersections, thus balancing the efficiency and safety of vehicles passing through unsignalized right-turn intersections.