36results about How to "Adapt quickly" patented technology

The invention discloses a multi-energy complementary power distribution control method, belongs to the technical field of electric power system control, and aims to solve the problems that an existing multi-energy system control method based on periodic instruction issuing is heavy in communication burden, response lags when sudden disturbance is dealt with in two scheduling intervals, and the power distribution efficiency is poor. And the operation economy and the dynamic stability are difficult to consider at the same time. The method comprises the steps that an upper-layer optimization controller generates virtual impedance potential field function parameters and key topological characteristics and issues the virtual impedance potential field function parameters and the key topological characteristics to a local controller; the local controller executes power control based on the potential field gradient and monitors the topological position of the operation point; when a topology crossing event is detected, an asynchronous event triggering signal is sent upwards, and an upper-layer controller is triggered to carry out aperiodic re-planning. According to the method, decoupling of economic optimization and dynamic control is realized, the robustness and the adaptive capacity of the system under disturbance are improved, meanwhile, the communication burden is reduced, and the method is suitable for a system containing a large amount of distributed energy.

The invention relates to the technical field of micro-grid load prediction, and discloses a multi-factor efficient fusion micro-grid load short-term prediction model optimization method. According to the method, micro-grid load historical data and multi-source factor data including information such as meteorology and electricity price are collected, and one-hot coding and normalization processing are carried out. Load data is decomposed into high-frequency and low-frequency components by using a variational mode decomposition technology, and prediction is carried out by respectively adopting a long-short-term memory network integrated with an attention mechanism and a linear regression model. And by calculating the difference between the source domain data and the target domain data in the model hidden layer feature space, the domain adaptive capacity of the model is enhanced, and accurate short-term prediction of the microgrid load is realized.

The high-strength three-way welding connection structure for the hydrogen conveying pipeline belongs to the field of three-way welding and comprises a base, two sliding grooves are formed in the upper surface of the base, first motors are fixed to the left side and the right side of the base correspondingly, and output shafts of the first motors penetrate into the sliding grooves and are fixedly connected with first screw rods; vertical frames are in threaded connection with the outer surfaces of the first screws, second motors are fixed to the opposite sides of the two vertical frames, and the opposite ends of output shafts of the two second motors penetrate to the opposite sides of the two vertical frames correspondingly. Supporting is provided from the inner wall of the three-way pipe through the supporting rollers, the transverse pipe positioning mechanism applies clamping force from the outer wall through the clamping blocks, a double-positioning structure of inner supporting and outer clamping is formed, it is ensured that three ports of the three-way pipe are strictly coaxial, and welding deflection is avoided; the clamping force can be automatically adjusted according to the small difference of the pipe diameters of the three-way pipes, the three-way pipes of different specifications can be rapidly matched through linear driving of the sliding frame and the air cylinder, and the universality is remarkably improved.

The invention relates to a rapid charge supply device for spacecraft pulse charge emission, and belongs to the field of space charging and discharging. The bottom plate is of a plate-shaped structure which is horizontally arranged in the axial direction; the cathode is of a cylindrical structure; the cathodes are in coaxial butt joint; the anode is of a cylinder structure; the anode is coaxially arranged in the inner cavity of the cathode; the root of the anode is butted with the bottom plate; the working medium sleeves the outer wall of the middle part of the anode and is filled between the cathode and the anode; the spring is sleeved on the outer wall of the anode root; one end of the spring is contacted with the bottom plate; one end of the spring contacts with the working medium end face; the insulating assembly is arranged at the opening of the axial head end of the cathode; two poles of the energy compression module are respectively connected with the cathode and the anode; the power supply is connected with the energy compression module to charge the energy compression module; a charge self-consistent cavity is formed on the outer side of the opening of the axial head end of the cathode; when the device is used for spacecraft pulse charge beam output, rapid supply in the pulse is achieved, and the spacecraft is kept electrically neutral in real time.

The application discloses a three-level coordinated production line body intelligent control method and system, and belongs to the technical field of intelligent manufacturing. The method comprises the following steps: establishing a global task planning layer, performing semantic analysis and logical disintegration at a first preset frequency, and generating an atomic action sequence; establishing a local space perception layer, receiving the atomic action sequence at a second preset frequency, constructing a dynamic local map by using sensor data, determining a workpiece pose by point cloud registration, and mapping the atomic action sequence into a continuous motion trajectory flow; establishing a real-time action execution layer, receiving an instantaneous target pose at a third preset frequency, calculating a theoretical driving torque by combining a Lagrange dynamics equation, switching to an impedance control mode at a contact moment, and driving an execution mechanism to complete work; and establishing an asynchronous coordination layer, realizing three-layer asynchronous communication through an industrial network, and controlling the three-layer frequencies to be distributed in a ladder type in an increasing manner. The application has the effects of improving the flexibility, precision and safety of a production line.

The utility model relates to the technical field of metal processing auxiliary equipment, and discloses an aligning and positioning tool for laminated processing of metal plates, which comprises a square guide rail, a sliding guide rail and a positioning component. A protruding block is arranged on the inner side of the square guide rail and connected with the two sliding guide rails in a sliding mode, a vertically-sliding positioning piece is arranged in each sliding guide rail, each positioning piece is composed of a sliding part and a rotating part, and the rotating parts are in rolling contact with the plate through the limiting holes. An elastic telescopic rod is arranged between the sliding guide rails to achieve self-adaptive adjustment, and the square guide rails drive the sliding guide rails to move oppositely through electric push rods. The working panel is provided with a guide groove and a slidable positioning sleeve, and plates are diagonally fixed through bolts. According to the tool, the electric push rod drives the positioning piece to conduct multi-direction linkage, the elastic telescopic rod and the rolling contact structure are combined, four-direction automatic positioning of the plate is achieved, the problems that a traditional tool is low in positioning efficiency and poor in precision are effectively solved, and the tool has the advantages of being capable of rapidly adapting to plates of different sizes, stable in multiple angles, capable of preventing deviation and capable of improving the machining precision.

The utility model discloses an unmanned aerial vehicle airborne equipment self-adaptive adjusting and quick dismounting device, which comprises a radar mounting plate, an airborne computer carrying plate, a clamping mechanism and an angle adjusting mechanism, and is characterized in that the clamping mechanism is arranged at the bottom of the airborne computer carrying plate and is used for clamping and separating from an unmanned aerial vehicle; a first stand column set and a second stand column set are arranged on the airborne computer carrying plate, a rotating shaft is arranged on the first stand column set, one end of the radar mounting plate is arranged on the first stand column set through the rotating shaft, and the other end of the radar mounting plate is arranged on the second stand column set through an angle adjusting mechanism. The angle adjusting mechanism is used for adjusting and locking the pitching angle of the radar mounting plate, and a placing space for placing an airborne computer is formed between the radar mounting plate and the airborne computer carrying plate. The device has the advantages of compact structure, high operation efficiency and capability of realizing real-time online processing of radar data.

The utility model discloses a motor vehicle maintenance creeper mechanism which comprises a creeper body, connecting rods are fixedly connected to the two sides of the creeper body, adjustable moving assemblies are installed on the two sides of the bottom of each connecting rod, and each moving assembly comprises an adjusting sleeve fixedly connected to the bottom of the corresponding connecting rod. Adjusting sliding rods are movably arranged on the inner walls of the adjusting sleeves, moving wheels are installed at the bottoms of the adjusting sliding rods, and positioning sliding blocks are fixedly connected to the tops of the adjusting sliding rods. By arranging the height-adjustable moving assembly, the height of the lying plate can be rapidly adjusted through simple operation, maintenance workers can rapidly adapt to chassis with different heights, adjustment time and physical output are reduced, the maintenance workers can work at a more comfortable position due to the copper wire height adjustment function, and the maintenance efficiency is improved. And the working efficiency and the working quality are improved, so that the use flexibility of the lying plate is improved.

The invention discloses a drainage basin ecological hydrological process simulation and prediction system based on a mechanism hydrological model and AI coupling, and relates to the technical field of drainage basin ecological hydrological simulation and prediction. Multi-source data integration and preprocessing are carried out through a data input layer, and a standardized data set is output and provided for a model coupling layer; the model coupling layer directly receives the standardized data set of the data input layer and outputs an optimized simulation result and a predicted value to the simulation and prediction layer; the simulation prediction layer receives a simulation result and a prediction value of the model coupling layer, performs multi-dimensional ecological hydrological element collaborative simulation and short-term-medium and long-term scene adaptive prediction, and outputs a simulation prediction result to the result output layer; and the result output layer receives the simulation prediction result of the simulation prediction layer and provides visual interaction and decision suggestion generation functions. The invention provides an innovative scheme of deep fusion of a mechanism model and an AI technology, and comprehensive requirements of physical rationality, data fitting precision, multi-element coverage and efficient response are covered.

This invention provides a training method and an evaluation method for a trajectory prediction model. The training method includes: obtaining the predicted trajectory of a target from the trajectory prediction model and dividing the predicted trajectory into multiple predicted sub-trajectory segments; obtaining the target's true trajectory and dividing the true trajectory into multiple true sub-trajectory segments, the number of true sub-trajectory segments corresponding to the number of predicted sub-trajectory segments; calculating the vector angles between the predicted and true sub-trajectory segments, including lateral and longitudinal vector angles; calculating the L2 distance loss between the predicted and true sub-trajectory segments; calculating the lateral and longitudinal distance losses based on the vector angles and L2 distance losses; and using the lateral and longitudinal distance losses as the loss function of the trajectory prediction model to optimize the parameters of the trajectory prediction model. This application performs a refined decomposition of the loss function, resulting in a more accurate prediction model after training.

The utility model relates to the technical field of positioning devices, and discloses an auxiliary positioning device of a full-automatic packing machine, which comprises a longitudinally arranged bearing plate, a sliding rail is arranged on one side wall of the bearing plate, a rack is arranged on the other side wall of the bearing plate, and the extending directions of the sliding rail and the rack are both transverse; moving plates are arranged at the two ends of the sliding rail in a sliding mode, bearing frames are arranged on the two moving plates, and guide baffles are arranged on the inner sides of the two bearing frames; and extension frames are arranged at the bottom ends of the two bearing frames correspondingly, through cooperation of driving motors and transmission gears on the two moving plates, the relative width between the two guide baffles can be automatically adjusted, and therefore the packaging requirements of objects of different sizes are met, and the flexibility and applicability of the packaging machine are improved. Due to the fact that the width of the guide baffle is adjustable, the guide baffle can rapidly adapt to objects of different sizes, the time for replacing the packaging template or manual adjustment is shortened, and therefore the efficiency of packaging operation is improved.

The utility model discloses an arc-shaped roof daylighting panel, and relates to the technical field of daylighting panels. Comprising a transverse telescopic plate; the number of the mounting plates is two, and the two mounting plates are arranged at the two ends of the transverse telescopic plate correspondingly; and the bottoms of the two sides of the arc-shaped plate are fixedly connected with the top of the mounting plate. Through the arrangement of the arc-shaped plate, the transverse telescopic plate and the mounting plates, the span and radian of the arc-shaped plate can be changed by adjusting the length of the transverse telescopic plate and changing the distance between the two mounting plates during use, the shape of the arc-shaped plate can be adjusted, and the arc-shaped plate can quickly adapt to an arc-shaped main body structure; and effective daylighting can be carried out on the interior of the arc-shaped roof.

The invention discloses a fast regional self-adaptive ACM method based on meta-learning, and belongs to the technical field of low-orbit satellite communication. Aiming at the technical problems of dependence on a large amount of local data, long debugging period, unstable performance and the like during cross-region deployment of the existing adaptive coding modulation technology, the method comprises the following steps of: constructing a meta-training task set covering various global climate characteristics, and training to obtain a meta-initial model with strong generalization ability; when a ground station is deployed in a new region, an optimized localized ACM strategy can be quickly adapted through several steps of gradient updating by using a very small amount of initial communication data collected by the station. Simulation results show that the method only needs 30 samples and 2-minute fine tuning to achieve the approximate optimal performance, the spectrum efficiency is improved by 35.8% compared with a traditional fixed threshold value method, the sample demand is reduced by 99% and the deployment time is shortened by more than 98% compared with a supervised learning method, the regional self-adaption problem in global rapid deployment of the satellite communication system is effectively solved, and the method has good application prospects. And the operation cost is obviously reduced.

The application relates to a multi-center applicable human brain magnetic resonance image brain structure automatic segmentation method, and the segmentation method specifically comprises the following steps: acquiring MRI data of different platforms, and performing N4 bias field correction and image standardization processing on the MRI data; taking U-Net as a basic framework, fusing a self-attention mechanism module, and constructing an MRI brain image structure segmentation network model suitable for multiple platforms; inputting the processed MRI data into the segmentation network model, training the segmentation network model by using transfer learning and domain adaptation technology, estimating the uncertainty of the output result, and further training the model by using high-quality unlabeled data screened out, so that the human brain magnetic resonance image brain structure automatic segmentation is realized. Compared with the prior art, the application can accurately and quickly segment the brain structure of different magnetic resonance imaging platforms, provides a more universal, efficient and reliable tool for the field of brain image analysis, and helps more extensive application and research exploration.

The utility model relates to silkworm cocoon processing equipment, and discloses a cocoon breaking machine capable of adjusting a paper cocooning frame. The equipment comprises a rack, a mounting frame and a cocoon breaking device, the mounting frame is provided with an adjusting frame, arc-shaped guide grooves are formed in the upper and lower surfaces of a front frame and a rear frame of the mounting frame, the mounting frame relatively slides along an arc-shaped track through a fixing rod, inner side connecting insertion pieces correct deformation of a paper cocooning frame, and an adjustable limiting block on the mounting frame limits the maximum sliding stroke of the frame. The limiting blocks are provided with kidney-shaped holes and scale marks, and adjustment and locking are convenient. The tail end arc structure of the insertion piece and the bottom conical structure of the cocoon breaking column reduce damage to the paper cocooning frame. The lifting driving mechanism is composed of a vertical guide rail, a lifting frame, a pedal plate, a transmission connecting rod and a reset spring, and lifting of the cocoon breaking device is achieved. The cocoon breaking machine is simple in structure, convenient to adjust and capable of efficiently adapting to paper cocooning frames of different specifications, damage is reduced, and the silkworm cocoon processing efficiency is improved.

The application discloses an adaptive intrusion detection method for an industrial internet, comprising the following steps: step S1: deploying an open set model on a global server and delivering parameters; step S2: training on a local data set by a client; step S3: uploading unknown features by the client, and performing cluster analysis on the unknown features by the global server to generate pseudo labels and corresponding initial class centers; step S4: performing rapid adaptive training based on a meta-learning mechanism by the client to obtain updated client local model parameters; step S5: aggregating group models by an edge server, and generating an updated model by the global server; step S6: iterating until convergence; and step S7: deploying the global model to perform real-time intrusion detection. The application can effectively perceive and adapt to learning new threats, protect data privacy in a distributed environment, break through the limitations of traditional models in a dynamic open environment, and improve the overall security and adaptability of the industrial internet system.

The invention provides a data classification method, which comprises the following steps: acquiring historical data, and screening the historical data by using an artificial intelligence AI small model to obtain important features of an object to which the historical data belongs; classifying objects to which the historical data belongs by utilizing an AI large model according to the classification description information and the important features; according to the embodiment of the invention, data classification is carried out by using a thought of cooperation of large and small models, complaint users and non-complaint users can be distinguished in a complaint user prediction scene, a large number of features with low complaint influence weights can be automatically screened out based on feature screening of an AI small model, the labor time consumption is reduced, the effectiveness of feature samples is also improved, and the accuracy of complaint prediction is improved. The prediction accuracy is improved; and meanwhile, rapid adaptation and efficient processing of the complaint prediction task can be realized by utilizing the generalization ability of the AI large model. The invention further provides electronic equipment, a readable medium and a computer program product.

The utility model discloses a seedling raising device for chive planting, which comprises a cultivation box and further comprises a lifting plate vertically arranged in the cultivation box, and the lifting plate can move up and down along the inner wall of the cultivation box to be adjusted; the partition plates are installed on the inner wall of the cultivation box in a sliding mode, the bottoms of the partition plates are attached and fixed to the top of the lifting plate, and the partition plates divide the upper portion of the lifting plate into a plurality of cultivation spaces; two extrusion plates are slidably mounted at the top of the lifting plate in each cultivation space, connecting rods are rotatably connected between the opposite surfaces of the two extrusion plates and the inner walls of the two sides of the cultivation box, and cultivation units are formed among the partition plates, the extrusion plates and the lifting plates. According to the chive seedling transplanting device, the soil balls are extruded before transplanting, so that root systems of chive seedlings can be tighter with the soil, the root systems are prevented from being separated from the soil, the root systems are protected, and the survival rate of the chive seedlings is remarkably improved.

The invention discloses an all-terrain anti-skid shoe sole which comprises a shoe sole body and at least one set of overturning components installed on the shoe sole body in an overturning mode, and the overturning components can overturn relative to the shoe sole body so as to adapt to different terrains. The overturning component comprises a supporting frame, protrusion matching parts and key bodies, the protrusion matching parts are arranged on the two sides of the supporting frame and used for achieving running fit, and the key bodies are in running fit with the protrusion matching parts. The overturning component which can be installed in an overturning mode is arranged on the shoe sole, so that the supporting frame can be rapidly overturned relative to the shoe sole; therefore, the shoe sole can rapidly adapt to various complex terrains such as ice and snow, mud and rocks, and potential safety hazards caused by poor terrain adaptability of an existing fixed type anti-skid shoe sole are avoided. And the technical contradictions of inconvenience in operation, insufficient durability and poor mechanism stability caused by structural homogenization of the existing reversible shoe sole are overcome.

The application discloses a three-dimensional target detection method and device based on language inference enhancement, comprising the following steps: collecting original images containing new categories from a current scene, generating diversified descriptions by using a multi-modal large model, obtaining images from the Internet, and performing data enhancement to construct a training set; training an open-world target detection model based on language inference; inputting the collected original images containing the target to be detected and the target category language prompt to be detected into the trained model to obtain a two-dimensional bounding box corresponding to the target category and a category confidence score in each original image; performing instance-level semantic segmentation on the image region of the two-dimensional detection box to obtain a target segmentation mask; performing coordinate conversion on the collected original three-dimensional point cloud of the target to be detected; and then associating the target segmentation mask with the projected laser radar point cloud to obtain a three-dimensional bounding box of the target to be detected. The application can quickly adapt to new categories and maintain strong generalization, and can significantly improve the detection accuracy of new categories.

This invention relates to the fields of image processing and artificial intelligence, and particularly to a few-shot object detection method based on support feature diffusion generation. The method includes steps such as constructing a few-shot object detection base class meta-learning training set and grouping image data within the training set. This invention introduces a support feature diffusion generation module, applying a diffusion model to the generation and optimization of support features. During the training phase, the inverse process of adding noise to features is learned, enabling the model to recover robust support feature representations from noise. During the testing phase, through progressive reverse diffusion, a more stable and diverse set of support features can be generated from a small number of support samples, effectively mitigating the feature instability and overfitting problems caused by few samples.

The application relates to the technical field of compound machine tools and discloses a variable working condition tool wear prediction method based on a compound machine tool, which comprises the following steps: s1: collecting dynamic signals and acquiring process parameters; s2: constructing a tool wear prediction model; s3: multi-task training of a meta-learning framework: through meta-learning, an updated W y And b y The tool wear prediction model is deployed in an actual machining system of the compound machine tool, dynamic signals and process parameters received during implementation are used, and a tool wear prediction value at a current time is quickly output. The application has the beneficial effects that the machine tool does not need to be stopped and the tool wear prediction accuracy is high.

The invention discloses a keratoconus element learning diagnosis method under an unbalanced small sample condition. The method comprises two parts of model training and auxiliary reasoning. According to the model training part, firstly, an acquired corneal topographic map is preprocessed and labeled, and then a meta-learning task ensuring complete categories is constructed by adopting a balance task sampler. And a pre-trained Swin Transform is used as a feature extractor to acquire the multi-scale depth features of the image. A plurality of prototypes are generated for early keratoconus categories with high diagnosis difficulty by using K-means clustering, and other categories are uniformly grouped. Classification is completed by calculating the distance between a query sample and a prototype and carrying out probability aggregation, and meta-learning optimization is carried out by adopting a joint loss function combining enhanced focus loss and contrast learning loss. And the auxiliary reasoning part is used for extracting input image features by using the trained model, comparing the input image features with the learned prototype, and outputting a diagnosis category and confidence. According to the method, the auxiliary diagnosis accuracy and generalization ability of the keratoconus are remarkably improved.

This utility model discloses an integrated set of equipment for excavation, loading, transportation, and unloading in long and deep foundation pit earthwork projects. It includes a tracked mobile bucket wheel reclaimer, a mobile telescopic conveyor, a prefabricated conveyor corridor, and a tracked mobile placing boom. After excavation, the soil from the tracked mobile bucket wheel reclaimer is fed into the inlet of the mobile telescopic conveyor. The outlet of the mobile telescopic conveyor is connected to the inlet of the prefabricated conveyor corridor, and the outlet of the prefabricated conveyor corridor is connected to the inlet hopper of the tracked mobile placing boom. The prefabricated conveyor corridor is supported by two tracked walking mechanisms near both ends. The tracked mobile placing boom is located outside the pit at the material stockpile site. This utility model eliminates the need for dump trucks for transportation, enabling efficient excavation and stockpiling of soil within deep foundation pits. Furthermore, it offers enhanced safety during transportation and stockpiling. All equipment can move via tracks, allowing for rapid relocation to the next location and improving relocation efficiency.

The invention provides a multi-robot online trajectory planning method and device and electronic equipment, and the method comprises the steps: respectively planning a global path of each robot in advance based on global obstacle information in a global map before each robot advances; and in the advancing process of each robot, sharing and obtaining the planning paths obtained by all the robots at the previous planning moment at each planning moment, and obtaining the planning path of each robot at the current planning moment under a plurality of set constraint conditions in combination with the planning paths and the global path. And if the phenomenon that the robot collides with other robots or is deadlocked possibly occurs, the planned path of the robot is adjusted based on the task priority. According to the scheme, real-time planning is executed in combination with the global planning path, and the planning path of each robot at the previous planning moment is shared, so that the phenomenon that the calculation amount is sharply increased due to the increase of the cluster number is avoided, the high real-time requirement can be met, the calculation overhead is reduced, and the control lag phenomenon is further avoided.

This utility model discloses a positioning mechanism for a PCB pin insertion machine, comprising: a base plate including a mounting through hole; a fixing block disposed above the mounting through hole, with its bottom ends connected to the base plate along a first direction, and its top end used to support the PCB board; a front limiting component and a rear limiting component respectively disposed on both sides of the fixing block, and symmetrically disposed side limiting components on both sides; and a driving unit disposed within the mounting through hole, including a telescopic cylinder and a telescopic connector, the driving end of the telescopic cylinder being connected to the rear limiting component; and both ends of the telescopic connector being connected to the side limiting components, used to drive the rear limiting component to move via the telescopic cylinder, and to drive the side limiting components to move via the telescopic connector, so as to abut and limit contact with the PCB board. Through the coordinated action of the front, rear, and side limiting components, multi-directional positioning can be achieved.

The application provides an oil-free mute air compressor tool, and particularly relates to the field of air compressor detection devices. The oil-free mute air compressor tool comprises a bottom plate, a bearing plate slidably connected to the bottom plate, a limiting component, a detection component and a driving component. The limiting component is provided with two groups on the bearing plate and is located at two ends of the bearing plate respectively. The limiting component is used for clamping an air compressor workpiece to perform testing. The oil-free mute air compressor tool provided by the application can switch work stations through the bearing plate, so that the testing process can be synchronized with the preparation process. The adjustable detection head and the limiting component can adapt to different models, so that the detection efficiency and the universality of the equipment are improved.

The utility model relates to electronic flower pattern machine technical field, specifically disclose a kind of electronic flower pattern machine and its template identification device, including electronic flower pattern machine body, clamping plate and push plate, clamping plate fixed mounting is in electronic flower pattern machine body front end bottom surface, clamping plate upper center position is equipped with clamping groove, push plate sliding installation is in clamping groove surface both sides, push plate surface both sides slidingly connected with slider, slider bottom fixedly connected with return spring, slider surface one side fixedly connected with pressing plate, push plate and pressing plate surface one side are all fixedly connected with hydraulic telescopic rod, pressing plate is lifted and adjusted on push plate surface by hydraulic telescopic rod, push plate is adjusted on clamping groove surface sliding by hydraulic telescopic rod, this electronic flower pattern machine and its template identification device, improve the versatility and applicability of equipment, can quickly adapt to different size and shape template, reduce the adjustment time caused by template replacement, improve the efficiency of template identification.

The present application belongs to the technical field of aero-engine thrust prediction, and particularly relates to an intelligent thrust prediction and real-time early warning method for an aero-engine. The specific technical solution is as follows: the test parameter characteristics of the aero-engine are clustered according to measuring components and systems; feature selection is performed on the clustered parameters; component and system sub-networks are established; the component and system sub-networks are connected according to the actual working matching relationship of the aero-engine; a feature mapping network is added at the final output end of the aero-engine digital engineering model; the network model is trained using test data generated in the test process of the aero-engine, and a real-time thrust prediction model and a benchmark thrust prediction model are obtained. The physical architecture of the aero-engine is integrated into the intelligent network design, efficient fusion of multiple systems in the digital space can be achieved, the digital model can closely track the performance of the aero-engine and reflect individual differences, and thus the engine thrust performance can be accurately tracked.

The invention relates to the technical field of automatic driving decision control and artificial intelligence, in particular to a multi-scene automatic driving strategy rapid adaptation method based on a meta-optimization framework, and the method comprises the steps: constructing a multi-scene expert strategy model based on reinforcement learning, and obtaining a multi-scene experience pool; constructing a meta-optimization fast adaptation model containing a representation learning network and a prediction learning network; meta parameters of the meta optimization fast adaptation model comprise representation learning network parameters and prediction learning network parameters; dividing the data of the multi-scene experience pool into a plurality of tasks, and performing cooperative training on the plurality of tasks based on a meta-optimization rapid adaptation model to obtain representation learning network parameters of the model; and when the automatic driving vehicle enters a new scene, keeping the representation learning network parameters of the model frozen, and performing multi-step gradient updating on the prediction learning network parameters by using a new sample. The objective of the invention is to solve the problems of weak strategy generalization ability and neural network plasticity loss of an automatic driving model in an unknown scene.