17results about How to "Reliable prediction" patented technology

The invention belongs to the technical field of software engineering, particularly relates to an automatic driving system performance enhancement method based on an expert hybrid architecture, and aims to solve the core problems that an end-to-end automatic driving system is confronted with semantic fuzziness to cause unreliable decision, multi-task interference hinders optimization planning, too long reasoning delay increases driving risks and the like. According to the method, an ExpertAD framework is provided, task key features are amplified through a perception adapter (PA), and the relevance of scene context understanding is guaranteed; related driving tasks are dynamically activated through a sparse expert mixture (MoSE), and task interference is minimized; and in combination with a customized training loss function, collaborative optimization of planning effectiveness and reasoning efficiency is realized. Experiments show that compared with an existing method, the method has the advantages that the average collision rate is reduced by 20%, the reasoning delay is reduced by 25%, higher multi-skill planning capacity is achieved in rare scenes (such as accident handling and first-aid vehicle avoiding), and good generalization is achieved for unseen urban environments.

The invention discloses an optical network routing and spectrum allocation algorithm based on a GCN-LSTM, and the algorithm comprises the steps: S1, defining an optical network environment, initializing the algorithm, and constructing a basic framework of a traffic prediction GCN-LSTM model; s2, according to historical traffic data in the network, network traffic data prediction is carried out in combination with a machine learning method; s3, predetermining an optimal modulation format based on the predicted flow result; and S4, performing efficient RSA decision according to the determined modulation format. According to the optical network routing and spectrum allocation algorithm based on the GCN-LSTM, optimization and improvement are carried out in the aspects of flow data prediction, flow dynamic threshold setting, modulation format determination and the like, and the problems that in a traditional heuristic algorithm, the spectrum resource utilization rate is low and the like can be solved.

This application provides a method, apparatus, and computer equipment for predicting the carbon dioxide diffusion coefficient in heavy oil enhanced oil recovery, relating to the field of oil and gas development technology. The method is based on Fick's law and the law of conservation of mass. It determines the saturated carbon dioxide concentration in heavy oil through carbon dioxide diffusion experiments. Simultaneously, it considers the volume expansion of heavy oil caused by carbon dioxide dissolving in it, and the oil relative flow opposite to the carbon dioxide diffusion direction caused by this volume expansion. Based on this, a partial differential equation is established to obtain a carbon dioxide diffusion model. The experimentally obtained saturated carbon dioxide concentration in heavy oil is used as the constraint condition for constructing the carbon dioxide diffusion model. The carbon dioxide diffusion model considers the volume expansion change of heavy oil as a one-dimensional expansion change, that is, the volume expansion change of heavy oil is considered through the change in the heavy oil liquid level. Finally, an iterative method is used to obtain a numerical solution. The above method achieves reliable and accurate prediction of the carbon dioxide diffusion coefficient.

The application discloses a battery pack energy state estimation method, device, equipment, storage medium and product, which are used for accurately and reliably estimating the energy state of a battery pack. The method comprises the following steps: acquiring a first operation data sequence of a first battery pack, wherein the first operation data sequence comprises operation data of the first battery pack at multiple time points; inputting the first operation data sequence into a first model to output a first energy state of the first battery pack; wherein the first model is obtained by training based on a second operation data sequence and a second energy state of a second battery pack, the single cell parameter distribution of the second battery pack is similar to the single cell parameter distribution of the first battery pack, and the single cell parameter distribution of the battery pack is obtained by a second model based on an operation data sequence of the battery pack.

The invention discloses an acoustic simulation method and system based on a low-frequency waveguide digital grid and geometric modeling, and relates to the technical field of acoustic simulation. A three-dimensional scene model composed of triangular patches is imported, and sound source parameters and the position of a receiver are initialized; based on the three-dimensional scene model, the sound source parameters and the receiver position, multiple sound field propagation paths are obtained through geometric acoustic pre-analysis, and initial perception importance parameters are calculated for each sound field propagation path. And distributing each sound field propagation path to a first calculation strategy or a second calculation strategy for processing according to the initial perception importance parameter. According to the method, through geometric acoustic pre-analysis and real-time dynamic evaluation, a high-precision calculation strategy is only allocated to a key sound field propagation path with high perceptual contribution degree, and an efficient geometric method is adopted for a large number of secondary paths. According to the resource scheduling strategy based on auditory perception, computing resources are highly focused, and meaningless consumption of secondary components is avoided.

This invention discloses a multimodal primitive learning-based traffic prediction method, system, device, and storage medium for anomaly events. The method includes: acquiring multi-source heterogeneous correlation data of the target road network, prior information of future events, and a static adjacency matrix; extracting a multimodal latent representation set, which includes at least a traffic latent representation; generating dynamic edge weight components based on the traffic latent representation and applying topological mask constraints using the static adjacency matrix to determine a time-varying adjacency matrix; performing graph propagation operations using the time-varying adjacency matrix to obtain enhanced spatial features, and aggregating these features with the remaining latent representations to obtain a fused node representation; converting the prior information of future events into a look-ahead conditional representation, merging it with the fused node representation for conditional decoding prediction, and outputting the future traffic state prediction result. This invention improves the model's spatial correlation capture capability and prediction accuracy under anomaly interference, achieving reliable prediction of traffic states in complex scenarios.

The application provides a marine clay-geogrid interface cyclic shear stress prediction model construction method, and relates to the technical field of marine engineering. The method comprises the following steps: preparing and assembling a marine clay-geogrid interface shear test sample; after the sample is assembled, a multi-working-condition interface cyclic shear test is carried out to obtain effective original test data; based on the effective original test data, an interface cyclic shear test database is constructed; based on the interface cyclic shear test database, a deep learning prediction model is constructed and trained; the accuracy of the prediction model is verified, the optimal prediction model is screened out, and a marine clay-geogrid interface cyclic shear stress prediction formula is built based on the optimal prediction model. The application can obtain reliable test data under multiple working conditions, accurately depict the shear stress evolution law by constructing a CNN-BiLSTM model, build an engineering usable prediction formula, and provide reliable support for the design and safety evaluation of marine engineering reinforced structures.

This application discloses a method and system for predicting the outlet moisture content of tobacco shreds during the drying process based on a KAN network, relating to the field of tobacco technology. This method anchors key process parameters associated with the outlet moisture content of dried tobacco shreds, avoids interference from redundant variables, and adapts to the strongly coupled multi-variable operating conditions of tobacco drying. Specifically, step S1 completes the screening of key process variables to ensure a strong correlation between input features and the prediction target; step S2 enhances the robustness of the model under operating conditions through standardized processing adapted to the characteristics of industrial data; step S3 constructs a network structure that overcomes the nonlinear expression limitations of traditional fixed activation functions; step S4 optimizes model parameters to improve generalization performance and interpretability; step S5 verifies the model's generalization ability to ensure prediction stability under unseen operating conditions; and step S6 achieves real-time prediction of outlet moisture content, compensating for detection lag. Overall, it achieves high-precision and high-reliability prediction of the outlet moisture content of dried tobacco shreds, balancing prediction performance and industrial application value.

The application discloses a multi-joint heavy hydraulic mechanical arm positioning error compensation method and system, and the method comprises the following steps: constructing a physical characteristic vector reflecting configuration, coupling and load; establishing an end X, Y and Z direction positioning error prediction model based on Gaussian process regression; predicting the current error by using the model, and compensating the target pose to generate an instruction pose to drive the mechanical arm to move. In the preferred embodiment, the error is decomposed into an explicit physical model prediction value and a GPR residual prediction value. The application combines physical priori and data driving, realizes high-precision error compensation under a small sample, significantly improves the positioning precision and generalization ability, and reduces the calibration cost.

The application relates to a high-frequency transformer core loss calculation method considering skin effect, and comprises the following steps: obtaining a magnetic hysteresis loop, calculating the average magnetic flux density and the magnetic field intensity in a period; performing Fourier analysis on the average magnetic flux density, calculating the eddy current magnetic field intensity corresponding to each harmonic, and obtaining the total eddy current magnetic field intensity after superposition; constructing a dynamic J-A magnetic hysteresis model, substituting the total eddy current magnetic field intensity to obtain an improved dynamic J-A model, and identifying the parameters in the improved dynamic J-A model; simulating the magnetic hysteresis loop based on the improved dynamic J-A model, and calculating the unit volume core loss. The application introduces the skin effect and realizes the collaborative modeling of the magnetic hysteresis and the eddy current, significantly improves the authenticity of the magnetic hysteresis loop simulation and the accuracy of the core loss calculation, thereby realizing the reliable prediction of the core loss under the high-frequency non-sinusoidal excitation working condition, and providing an analysis method with more engineering value for the optimized design and performance evaluation of the high-frequency transformer.

The application discloses a load aggregation body prediction method and system of an integrated framework, and the method comprises the following steps: applying a prediction result of a trained load aggregation body prediction model to perform scheduling; wherein the construction of the load aggregation body prediction model comprises the following steps: obtaining load aggregation body historical data; performing external influence factor analysis on the load aggregation body historical data, and screening important external influence factors; applying a K-shape algorithm to cluster the load aggregation body historical data, and dividing the load aggregation body historical data into a plurality of clusters; constructing a load aggregation body prediction model, including a CNN-LSTM-ATTENTION prediction model and a GBDT prediction model; performing load prediction on each cluster through the CNN-LSTM-ATTENTION prediction model by using the screened important external influence factors and the divided plurality of clusters; and applying the GBDT prediction model to perform integrated prediction on the load prediction results of the clusters, thereby obtaining a load aggregation body prediction result; and the application can improve the prediction accuracy of the load aggregation body prediction model.

The invention provides a visual tracking prediction method and device for road traffic and a storage medium, and relates to the technical field of trajectory prediction. The visual tracking prediction of the road traffic comprises the following steps: obtaining multiple frames of first observation images collected by a first camera before a target vehicle enters a blind area; obtaining road structure information of a road on which the target vehicle travels; searching a target driving track matched with the target vehicle from a plurality of preset driving tracks of a blind area in a historical database; and predicting the driving track of the target vehicle in the blind area according to the multiple frames of first observation images, the road structure information and the target driving track. According to the method, information of multiple dimensions, such as multiple frames of first observation images before the target vehicle enters the blind area, road structure information of a driving road, a matched target driving track in a historical database and the like, is fused, the actual traffic law is better met, and the driving track of the target vehicle in the blind area can be accurately and reliably predicted based on the information.

The application relates to the technical field of battery early warning and discloses a battery alarm scene data enhancement and algorithm robustness verification method, which acquires a real alarm sample set; a pre-trained cascade generation model is used to generate a synthetic alarm sample; the synthetic alarm sample and the real alarm sample are mixed in proportion to construct an enhanced training set; a battery alarm related target recognition algorithm is trained using the enhanced training set, and the robustness improvement degree of the battery alarm related target recognition algorithm is quantitatively verified under a plurality of preset disturbance conditions; wherein the cascade generation model comprises the following steps: a denoising diffusion probability model is used to generate a basic sample, a generative adversarial network is used to perform detail enhancement and distribution correction on the generated result, a preset strategy is used to control a random latent variable to complete sample screening, and a high-quality and diversified alarm data set is constructed. Through the method, the robustness and stability of the alarm recognition algorithm can be significantly improved, and reliable prediction and accurate early warning of a battery system under actual complex working conditions can be realized.

This invention discloses an edge contour determination method, a mask pattern correction method, an apparatus, and an electronic device. The edge contour determination method includes obtaining the lithographic intensity value of each grid pixel in an initial simulated lithographic pattern obtained by lithographic simulation of a mask pattern using a lithographic simulation model; determining edge contour points whose lithographic intensity values ​​are equal to a set intensity threshold based on a pre-determined intensity variation trend curve along the edge contour normal direction of the lithographic pattern and the corresponding lithographic intensity value of each grid pixel; and determining the edge contour curve of the initial simulated lithographic pattern based on each edge contour point. In this application, the intensity variation trend curve along the edge contour normal direction of the lithographic pattern is pre-determined, thereby achieving simpler and more accurate and reliable extraction of edge contour points, thus enabling accurate and reliable prediction of the lithographic pattern. When applied to mask pattern correction, this method is beneficial for improving the precision of semiconductor lithography.

The application provides a business processing method and device, which can be used in the financial field or other fields. The method comprises the following steps: receiving a business processing request of a target business scenario, wherein the business processing request comprises multiple types of business data; determining whether the business processing request is verified according to a preset switch rule decision tree and each type of business data, and if yes, completing the business processing corresponding to the business processing request; wherein the preset switch rule decision tree comprises rules of multiple business scenarios and corresponding relationships between the rules, and the corresponding relationships comprise and relationships. The application can guarantee the universality and reliability of the business switch, thereby improving the reliability of the business processing.

The application discloses a kind of life prediction method, device, equipment and medium of gas insulated switchgear, belong to gas insulated switchgear life prediction field.The method is: from gas insulated switchgear, gas sample is collected, and laser-induced breakdown spectroscopy technology is used to excite sample and obtain NO spectrum signal data;NO spectrum intensity is calculated according to preset wavelength parameter;Further, NO spectrum intensity is input into prediction model, to predict the switch breaking number of gas insulated switchgear and assess its life;Wherein, prediction model is obtained by establishing the quantitative relationship between NO spectrum intensity and switch breaking number.Therefore, by implementing the present application, the problem that it is difficult to ensure prediction accuracy while achieving convenient on-site life prediction in the prior art can be solved.

The invention provides a resource effect prediction method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining candidate structured configuration information and semantic description information of a target resource; according to the candidate structured configuration information, evaluating a target service index of the target resource under the candidate structured configuration information to obtain a first predicted value of the target service index; according to the semantic description information, evaluating a target business index of the target resource under the candidate structured configuration information to obtain a second predicted value of the target business index; and determining a target prediction result of the target business index according to the first prediction value and the second prediction value. According to the method, semantic features contained in numerical configuration parameters and unstructured text contexts are effectively fused, the comprehensiveness and accuracy of target business index prediction are improved, the reasonability of resource configuration decisions can be enhanced, and the pre-judgment ability and configuration efficiency of a system before resources are online can also be improved.