24results about How to "Verify robustness" patented technology

The invention discloses a multi-modal MRI (Magnetic Resonance Imaging) reconstruction method based on Transform, which comprises the following steps of: inputting paired under-sampling and full-sampling MRI image pairs into a pre-trained multi-modal MRI image reconstruction network model to obtain a high-resolution image after target modal reconstruction; wherein the multi-modal MRI image reconstruction network model comprises a spatial registration module and an image reconstruction module, the spatial registration module is used for carrying out spatial position alignment on an auxiliary modal MRI image to a target modal MRI image, and the image reconstruction module is used for extracting auxiliary information from an MRI image pair; the multi-modal MRI image reconstruction method is suitable for multi-modal MRI image reconstruction, and the detail quality of the reconstructed MRI image can be effectively improved.

The invention relates to the field of image or video recognition, in particular to a local learnable query enhancement video recognition method, which comprises the following steps of: establishing a frame for adding video space-time modeling and migrating to a video task on the basis of an image large model; constructing a video behavior recognition model in the framework; constructing a training mechanism of large model main branch freezing and only new parameter learning, training a video behavior recognition model by using a server, and optimizing an objective function until network convergence to obtain a trained video behavior recognition model; and inputting a video sequence to be recognized into the trained video behavior recognition model to recognize human behaviors. The method has the beneficial effects that a video behavior recognition task can be quickly and efficiently migrated by only adding a small number of learnable parameters on the basis of an image model, human behaviors in a video can be recognized with high precision, and the method has high robustness and effectiveness in the aspect of processing action related details.

The invention discloses a multi-loop natural circulation coupling heat removal experiment system, which relates to the technical field of nuclear reactor safety, and is characterized by comprising an integrated pressure vessel, an energy conversion and residual heat removal integrated design loop, a condenser cooling water loop, a thermal insulation material and a measuring instrument, the heat insulation material coats the heat extraction experiment system and is used for reducing energy leakage; and the measuring instrument is arranged on the heat extraction experiment system and is used for measuring and monitoring the flow, the temperature, the pressure and the liquid level. Research is carried out on an integrated pressure container designed in a marine nuclear power source and an energy conversion and waste heat removal integrated system, and three natural circulations are formed in a secondary loop and a condenser seawater section in the integrated pressure container respectively; all natural circulations are coupled into a whole through a steam generator heat transfer pipe and a condenser heat transfer pipe, and waste heat of a reactor core is discharged through coupling of multiple natural circulations, so that shutdown safety of a nuclear reactor is guaranteed.

The invention discloses a battery SOC prediction method fusing online identification and an adaptive algorithm, relates to the technical field of power battery monitoring, and aims to solve the problem of insufficient SOC prediction precision of an electric vehicle power battery under a dynamic working condition. The prediction method comprises the steps of obtaining lithium battery charging and discharging experiment data, establishing a dual-polarization equivalent circuit model to simulate a strong nonlinear behavior in the battery, and fitting an OCV-SOC curve based on the experiment data; performing online parameter identification on the equivalent circuit model by using a forgetting factor recursive least square method based on grid search-real-time voltage difference improvement; sOC prediction is carried out by combining online parameter identification with a self-adaptive expansion Kalman filtering algorithm; and the proposed prediction algorithm is verified under a mixed pulse power characteristic test working condition. According to the method, the accuracy and the real-time performance of SOC prediction are remarkably improved, and the method has a wide application prospect.

The present invention belongs to the field of control technology, and relates to an event-triggered robust nonlinear model predictive control method for agricultural machinery trajectory tracking. The method comprises: establishing a transverse and longitudinal dynamics model of agricultural machinery and its state space equations, and tracking the operation trajectory of agricultural machinery; constructing an improved nonlinear model predictive control system, the improved nonlinear model predictive control system comprises a nonlinear controller of the agricultural machinery nominal system with a pseudo-Huberl function improved cost function, and solving the control input of the agricultural machinery nominal system; optimizing the solver, and using the reinforcement learning actor-critic algorithm to approximate the uncertainty compensation of the agricultural machinery model; using an adaptive parameter nonlinear disturbance observer to estimate and compensate for the external disturbances suffered by the agricultural machinery; through an event triggering mechanism, judging whether to trigger the optimization solver to optimize the solution according to the error between the real value and the predicted value of the agricultural machinery trajectory tracking, and obtaining the input of the agricultural machinery trajectory tracking controller. The present invention enhances the robustness of the control system and reduces the calculation frequency.

The invention relates to a method and a device for automatically testing SPI peripherals, which are based on a designed upper computer and lower computer test architecture and test mode, the test flexibility is high, parameters can be customized according to requirements, and various SPI signals can be flexibly generated; the test is more comprehensive, random data is supported to better verify the pressure resistance and robustness of the SPI peripherals, potential boundary conditions and abnormal conditions can be found, and the stability and reliability of a vehicle-mounted system are ensured; the application range is wide, the reusability is high, and the MCU with SPI (Serial Peripheral Interface) equipment can be tested; the test operation is simple, the test efficiency is high, the cost is low, and a single chip can be self-tested to reduce interference.

This invention discloses an adaptive proximity valve block assembly and its control method based on hydraulic potential energy recovery. The valve block assembly is fixedly connected to a hydraulic cylinder. The valve block internally includes a balance and stabilization passage, a pre-charge buffer passage, a flow-limiting access passage, a main recovery passage, and a buffer chamber. The pre-charge buffer passage has a one-way throttling element and connects to the buffer chamber. The flow-limiting access passage has a fixed throttling orifice connecting the buffer chamber and an energy storage unit. The main recovery passage has a hydraulically controlled directional valve driven by the pressure difference between the buffer chamber and the energy storage unit. After recovery is enabled, the high-pressure return oil preferentially enters the buffer chamber to absorb the first wave of pressure redistribution due to the fast-forward characteristic of the one-way throttling element. Then, it slowly releases the second wave of impact by charging the energy storage unit at a low flow rate through the fixed throttling orifice. After the pressure difference shrinks to a threshold, the hydraulically controlled directional valve automatically opens the main recovery passage. This invention reconstructs the recovery entry from a one-step opening to a hydraulically automatic graded access sequence, significantly suppressing entry pulsation, avoiding main control misjudgment, and improving execution stability.

This invention discloses a method for registering industrial printed images based on a lightweight network. The method includes: S1 Image acquisition and dataset construction, which uses a CCD industrial camera to acquire images of the front and back of printed materials and constructs a training dataset by combining multispectral images; S2 Constructing a DHSiam network, which includes two parallel feature extraction networks; S3 Proposing an attention heterogeneous mechanism, which includes a position regression unit and a lightweight unit to enhance data representation and reduce model parameters; S4 Constructing a cross-entropy loss function with additional margin to enhance network performance metrics. This invention builds a hardware platform with 8 parallel GPUs, uses only AMC-Loss to distinguish image similarity, utilizes the attention heterogeneous mechanism to extract important information from feature maps at different scales, and adjusts model parameters to enhance the algorithm's performance on mobile devices, thereby comprehensively improving the model's ability to register printed images with different perspectives and deformations.

The application discloses a kind of multi-agent preset time formation control methods based on backstepping method under distance rigidity matrix, comprising: 1) based on the double integrator dynamics model of multi-agent system is established based on rigidity graph theory, define the minimum and infinitesimal rigidity frame corresponding to the expected formation;2) introduce backstepping design framework, define virtual speed error variable;3) introduce time-varying scaling function to construct time scale transformation;4) design distributed preset time controller;5) through numerical simulation and physical platform are verified, verify the effectiveness of distance error in user preset time under different initial conditions converges to the vicinity of zero.The application is fused by backstepping and preset time scaling function, for the first time in only local relative measurement distance rigidity formation Convergence time user self-defined preset is realized, and convergence time upper bound is independent of initial condition and controller parameter, provides reliable, engineering implementation solution for time-critical multi-agent cooperative task.

The present invention discloses an electromagnetic signal feature intelligent classification method based on lightweight progressively differentiable architecture search, comprising the following steps: step 1, using the pre-processed IQ samples generated by MATLAB as a proxy data set; step 2, constructing a search space and a unit structure to search for normal units and reduced units; step 3, searching for the best network architecture using a progressively differentiable architecture search method based on step 2; step 4, the training process of the final architecture; step 5, migrating the searched structure to the target task RadioML2016.10B; step 6, retraining to obtain the final model. The method disclosed by the present invention considers the classification accuracy and computational complexity of the model at the same time, achieves a balance between performance and computational overhead, and reduces the complexity of the model without almost reducing the modulation classification accuracy.

The invention aims to provide a phase unwrapping method based on extended Kalman filtering. The method comprises the following steps: A, constructing a two-dimensional phase unwrapping program MCC-EKFPU; and B, performing phase unwrapping processing on the interferogram by using a two-dimensional phase unwrapping program MCC-EKFPU to obtain an unwrapped phase of the interferogram. According to the method, the maximum correlation entropy criterion extended Kalman filtering theory is introduced into interferogram wrapping phase unwrapping, and an interferogram phase gradient estimation technology and a robust path tracking strategy are combined, so that the method has higher phase unwrapping precision and better robustness; the method has good application potential in interferogram phase unwrapping application with high accuracy requirements.

The invention discloses an experimental verification method of a multi-loop natural circulation coupling waste heat removal system, and relates to the technical field of nuclear reactor safety, and the technical scheme is characterized in that the method comprises the following steps: optimizing size parameters and thermal hydraulic parameters of a structure and equipment in the overall preliminary design of the system; a bidirectional multi-stage modeling method is adopted, and a multi-loop natural circulation coupling residual heat removal experiment system is built; and carrying out operation characteristic research and influence analysis on the multi-loop natural circulation coupling residual heat removal experiment system, and verifying the operation feasibility and robustness of the multi-loop natural circulation coupling residual heat removal experiment system. Research is carried out on an integrated system of a marine nuclear power source, three natural circulations can be formed in an integrated pressure container, a secondary loop and a seawater section of a condenser respectively, all the natural circulations are coupled into a whole through a steam generator heat transfer pipe and a condenser heat transfer pipe, reactor core waste heat is discharged through coupling of the multiple natural circulations, and the heat transfer efficiency is improved. And the shutdown safety of the nuclear reactor can be guaranteed.

The invention provides a circuit model based on a one-dimensional Su-Schriffer-Heeger lattice structure. The circuit model comprises a circuit model for carrying out lattice structure building through a typical SSH model; the crystal lattice comprises two non-equivalent lattice points; the two non-equivalent grid points are connected in a specific mode and simulate different transition intensities, and each grid point is regarded as a node in the circuit. According to the invention, the appearance of the topological end state in the one-dimensional structure is verified by means of theoretical calculation and experiments, the influence on the end state characteristics is analyzed by introducing different defects into the end state characteristics existing in a circuit lattice system, and the end state characteristics are verified to have certain robustness. Through simulation of the topology end state phenomenon in a circuit system, it can be predicted that the topology end state phenomenon possibly exists in the structure system in a condensed state, and development of novel electronic materials is greatly facilitated through research of the topology end state phenomenon in the condensed state system.

The present invention discloses an electrical fault diagnosis method for a permanent magnet synchronous motor under different working conditions and different noise environments. The zero-sequence voltage component and the three-phase stator current are used as network inputs to construct a multi-resolution convolutional network, extract multi-layer perception information, and realize fault discrimination. Simulation data containing 34 states are generated using simulation software to construct a simulation data set. Experimental data of 34 states under real experimental conditions are collected to construct a real data set; the simulation data set and the real data set are divided into three subsets: a training set, a verification set, and a test set. A multi-resolution network is built to extract fault features to realize fault classification; Gaussian noise and Laplace noise are added to the data to verify the robustness of the network to noise. The present invention realizes for the first time the use of a deep learning method to simultaneously diagnose voltage source inverter open circuit faults, stator winding open circuit faults, and turn-to-turn short circuit faults under different working conditions and different noise environments.

This invention discloses a method and system for collaborative scheduling of multiple types of V2G resources in the electricity-carbon certificate coupled market, belonging to the field of resource collaborative scheduling technology. The method includes: designing a two-way conversion mechanism between carbon quotas and green certificates in an established first model, introducing a carbon emission factor; constructing a master-slave game and Wasserstein split-bar optimization model based on the first model with the two-way conversion mechanism; establishing objective functions for a DSO multi-market joint clearing model and an EVA split-bar Mean-CVaR bidding decision model respectively; solving each objective function based on the constraints corresponding to each model, and performing collaborative scheduling of resources based on the solution results; balancing the economy and robustness of the scheduling strategy, achieving collaborative response of multiple types of V2G resources in the electricity-carbon-certificate multi-market, effectively improving the system's operational economy, power supply reliability, and decarbonization level.

The invention discloses a single-phase earth fault detection method based on staggered phase deformation and a first half wave, and the method comprises the steps: starting from the deformation angles of a zero-voltage signal and a zero-current signal after an earth fault occurs, carrying out the staggered phase deformation detection of the collected zero-voltage signal and zero-current signal, determining the fault occurrence time, and carrying out the judgment under different faults. Furthermore, a first half-wave method is combined to realize efficient diagnosis of the fault section of the medium-voltage power distribution network, and the feasibility, effectiveness and robustness of the technical scheme of the invention are verified through specific embodiments; the defects of an existing traditional single-phase earth fault detection method in diagnosis of various single-phase earth faults (such as gradual change faults, intermittent faults and successive faults) are overcome. Besides, the new method provided by the invention has another advantage that frequency domain or power energy conversion does not need to be carried out, hardware and example requirements are not increased, and the method can be directly embedded into the existing detection device.

The invention provides a double-station multi-target tracking method and device based on WMFS and UKF, and the method comprises the steps: constructing a WMFS model, and reconstructing a state transfer function through a pre-trained WMFS model; combining the optimized WMFS model with the UKF model to obtain a double-station multi-target tracking model; and tracking a target by using the double-station multi-target tracking model. According to the method, the advantages of the WMFS in the aspects of processing system uncertainty and complex modeling are fully utilized, in the WMFS-UKF derivation process, an unscented (UT) frame is adopted to solve the nonlinear measurement problem, meanwhile, a pre-trained WMFS is used for reconstructing a state conversion function, and through the historical state and the expected output of a target, the nonlinear measurement problem is solved. The WM fuzzy inference system realizes more accurate state prediction and accurate covariance estimation, improves the performance and stability of target tracking, and verifies the effectiveness and robustness of the algorithm in a target tracking scene.

A method and system for rapid simulation of a spacecraft system based on an image model closed loop, comprising: establishing a three-dimensional digital model of a spacecraft, wherein the size of the spacecraft is adjustable and the characteristic load of the spacecraft is configurable; setting the space environment and environmental information of the simulation scene; retrieving the observation area according to the sensor installation configuration to obtain an ideal image of the observation area; establishing an image sensor model to simulate the imaging effect of the optical system and degrade the ideal image of the observation area; degrading the optical simulation image by the circuit system to obtain an observed simulated degraded image; processing the observed simulated degraded image; simulating the digital interface of the imaging sensor, saving and generating the simulation image, and saving or transmitting the data that meets the electrical characteristics and timing characteristics of the sensor electronic circuit interface to a subsequent processing unit; connecting the simulation closed loop of the spacecraft control system through a standardized data interface, and solving, storing and displaying the simulation results through the master control scheduling.

The application discloses a kind of lithium ion battery distributed thermal process sensor fault estimation method, comprising: using two-dimensional partial differential equation to describe rectangular lithium ion battery distributed thermodynamic equation;Distributed thermal model of large size lithium ion battery is constructed;Using Chebyshev-Galerkin method, distributed thermodynamic equation and distributed thermal model are decomposed into reduced order model described by standard state space equation in time domain;State variable in reduced order model is constructed using sensor measurement output and a Hurwitz matrix, and enhanced reduced order model is constructed;Enhanced adaptive observer and error state space equation are constructed using enhanced reduced order model;Based on error state space equation, the fault strength of temperature sensor that fails is estimated using fast adaptive algorithm.The application can effectively estimate sensor fault, including time-invariant and time-varying fault, and single sensor and multiple sensor fault.

The present invention discloses a global optimization method for decomposing and executing complex data query tasks, which is based on a tree structure and is implemented using a large language model, including constructing a task decomposition and execution module based on a tree structure: the task tree adopts a tree structure, the root node represents the overall data query task, each child node represents a subtask decomposed from a parent node, and the state of each branch corresponds to the execution state of the corresponding subtask; constructing a task execution self-correction module based on a tree structure: when a subtask fails to execute, quickly locate the faulty child node and the corresponding parent node, and re-execute the faulty node; a lightweight local database knowledge graph module based on a local database and oriented to the basic attributes of the database: extract key information from the table structure of the database, and convert it into a knowledge graph through information combination; the present invention has significant effects and superior performance in processing complex database queries.

The invention belongs to the field of automatic control technology and multi-agent system cooperative control, and particularly relates to a multi-agent system fault-tolerant inclusion control method and system based on a neural network, and the control method comprises the following steps: S1, constructing a second-order multi-agent system; s2, constructing a full connection-fuzzy neural network; performing high-precision online fitting and approximation on non-ideal factors in the second-order multi-agent system to generate corresponding compensation items; s3, constructing a fixed time state observer, and estimating unknown position states and speed states of all the following intelligent agents within a preset time upper limit irrelevant to the initial state of the system; s4, constructing a fixed time sliding mode control law based on the estimation state and the compensation item; and S5, applying a fixed time sliding mode control law to the following agents, so that all the following agents converge and are stabilized in a dynamic convex hull formed by a plurality of leader states within a fixed time, and realizing inclusion control. According to the invention, the risk in practical application can be effectively reduced.

The present invention discloses a method and device for equipment fault diagnosis based on a multi-scale 2D dilated convolutional network, and relates to the technical field of mechanical equipment fault diagnosis. The method includes: obtaining mechanical equipment data; converting the mechanical equipment data into a two-dimensional image; inputting the two-dimensional image into a trained multi-scale 2D dilated convolutional neural network; and obtaining mechanical equipment fault diagnosis results based on the two-dimensional image and the trained multi-scale 2D dilated convolutional neural network. The present invention introduces dilated convolution into the network, flexibly realizes multi-scale feature extraction, reduces information redundancy, reduces calculation amount, and reduces the number of model parameters; introduces a cascade structure into the network, combines low-resolution information with high-resolution information to improve the fault diagnosis accuracy when the number of samples is insufficient, and improves the model accuracy when the samples are noisy.

The invention discloses a multi-natural circulation coupling heat extraction device for an ocean nuclear power source, which relates to the technical field of nuclear reactor safety, and is characterized by comprising an integrated pressure vessel structure, an energy conversion and residual heat removal integrated design loop structure and a condenser cooling water loop structure, the energy conversion and waste heat removal integrated design loop structure is connected with the integrated pressure container structure and the condenser cooling water loop structure. Research is carried out on an integrated pressure container designed in a marine nuclear power source and an energy conversion and waste heat removal integrated system, and three natural circulations are formed in a secondary loop and a condenser seawater section in the integrated pressure container respectively; all natural circulations are coupled into a whole through a steam generator heat transfer pipe and a condenser heat transfer pipe, and waste heat of a reactor core is discharged through coupling of multiple natural circulations, so that shutdown safety of a nuclear reactor is guaranteed.

This invention belongs to the field of knee joint MRI segmentation technology, specifically relating to an efficient knee joint 3D MRI segmentation method and system based on iterative pseudo-label refinement. The method includes: constructing a training set based on knee joint 3D MRI volume data; training an initial segmentation model using a pseudo-label self-refinement strategy based on the training set to obtain a refined iterative segmentation model; and implementing knee joint 3D MRI segmentation based on the refined iterative segmentation model. This invention first analyzes the relationship between different annotation ratios and segmentation performance, and based on this, proposes an iterative pseudo-label refinement strategy. By progressively optimizing the quality of pseudo-labels to reduce reliance on manually labeled data, a more efficient knee joint MRI segmentation method is explored.