40results about How to "Reasonable description" patented technology

The invention discloses a sign language recognition method and system based on a double-flow space-time diagram convolutional neural network, and the method comprises the steps: firstly segmenting a sign language motion video into video frames, extracting the upper body and hand skeleton points of a person in a sign language motion video segment, and constructing global and local diagram data; respectively extracting global and local spatial-temporal features by utilizing a double-flow spatial-temporal graph volume network, and obtaining global-local features through feature splicing; meanwhile, texts corresponding to the videos are encoded into word vectors after word segmentation processing, the word vectors and the texts are mapped to the same implicit space through feature transformation, and model training is conducted through a dynamic time warping algorithm; and for the global-local feature sequence, a self-attention mechanism coding and decoding network is adopted to perform serialized modeling on the global-local feature sequence, and a softmax classifier is adopted to obtain words corresponding to each video clip by the output of a decoder, and corresponding text sentences are formed. According to the method, the accuracy of text sentence generation can be improved, and the method has important application value in scenes such as caption generation and human-computerinteraction.

The invention discloses an improved fuzzy neural network bus intelligent scheduling method based on a chaos theory, and belongs to the field of intelligent transportation. According to the improved particle swarm bus intelligent scheduling method based on the chaos theory, advantages and complementarity of various algorithms are fully utilized, a series of improvement measures are also introduced, such as conjugate gradient optimization, and inertia factor and constraint factor of the particle swarm algorithm etc., the mechanism and the search performance are researched from the theoretical and practical perspectives, problems of poor global search capability and premature convergence of the conventional optimization algorithm are fundamentally solved, the diversity of population can be obviously increased, the global search capability is obviously improved, the problem of fuzzy information can be effectively dealt with, the convergence speed is fast, and a new high-efficiency method is provided for bus intelligent scheduling.

This invention is based on the time and events sequence to analysis attacking detection methods and devices involved in exchange for the function of the network. It is a methods and devices to prevent removing data from the data transmission channel without permission .The method provides a complex text-based attack description language to make the user can amend the built-related characteristics, and add new features of related events. The present invention include: basic rules of the incident, the rules of the association, first level detection engine, correlation analysis engine. The whole process of attacking are described more comprehensive and reasonable taking into account the time factor and the order of events Such description and testing which distinct the basic incidents more carefully are more in line with the requirements of detecting attacks. The invention also describes the relation between alarm incident and not alarm incident.

The invention relates to a method for measuring the flexural rigidity of an immersed tube tunnel joint. The method comprises the following steps: 1, applying an axial external water pressure N0 to the immersed tube tunnel joint as an external force applied to a position having a distance of an initial eccentric throw e0 to the neutral axis of the immersed tube tunnel joint; 2, continuously applying prestressed cable pretensions T1 and T2 to the immersed tube tunnel joint; 3, measuring the displacement S0 and the corner theta of a GINA waterstop under the action of N0, T1 and T2; 4, using an immersed tube tunnel joint structure simplification model established by a steel plate and a spring element to calculate the compression amount SF1 and SF2 of the GINA waterstop, the prestressed cable extension amount ST1 and ST2 at a top board and a bottom board, the pressures F1 and F2 of the GINA waterstrop of the top board and the bottom board, and the pulling forces T1 and T2 of the prestressed cables of the bottom plate; and 4, calculating according to a force balancing equation to obtain the flexural rigidity Ktheta of the joint. Compared with the prior art, the method provided by the invention improves the design level and efficiency of the performance of the immersed tube tunnel joint.

The invention discloses a vehicle trajectory prediction method and system based on video vehicle passing data. The method comprises the following steps: collecting vehicle passing data, and extractingeffective field information in the vehicle passing data; constructing a basic road network; sorting the vehicle passing data according to the sampling time to form a total travel chain of each vehicle in the sampling time period; extracting two adjacent sample sequences in turn, calculating the travel time of the vehicle passing through two adjacent intersections according to the sample sequences, judging that the two adjacent intersections are travel chain segmentation points, and segmenting the total travel chain into a series of sub-travel chains according to the travel chain segmentationpoints; calculating the steering state probability of each vehicle at each intersection according to the sub-travel chain to obtain a steering probability matrix; determining that the intersection corresponding to the maximum steering probability in the steering probability matrix is the intersection to be reached by the vehicle driving trajectory prediction. The invention avoids the shortcomingsof relying on continuous positioning data, so as to realize the travel behavior characteristic analysis of any vehicle.

The invention relates to a method for evaluating the degree of blasting damage of surrounding rock in an underground water-sealed cavern. A set of blasting vibration monitoring points is arranged behind the tunnel face; the naked rock excavated in the excavation is classified based on the Q classification method, and the equivalent rock mass A value of a group of monitoring points is obtained through formula conversion; the law of propagation of surrounding rock vibration is obtained by correcting and fitting the Sadowssky formula using the maximum single-section charge amount, the distance from the burst center, the combined peak speed and the A value of one blasting; the prediction value of the combined peak vibration speed is obtained based on the law formula of propagation of surrounding rock vibration and follow-up blasting parameters and the Ai value of each monitoring point, and is compared with the measured value to obtain the vibration speed reduction rate of each monitoring point, and the arithmetic mean value is taken as the degree of blasting damage of surrounding rock. The method takes into consideration the influence of the rock mass on the propagation of surrounding rock blasting vibration. It is reasonable and reliable to use the arithmetical mean value of the combined peak vibration speed reduction rate of each monitoring point as the evaluation index of the damage degree of the surrounding rock.

The invention provides a semantic network object identification and judgment method. The semantic network object identification and judgment method aims to provide a comprehensive judgment method capable of lowering calculated amount of evidence compound and update, lowering confliction among evidences with multiple attributes, and improving consistency of identification results. The method is achieved through the following technical scheme that a target semantic database, an evidence receiving module, an evidence semantic knowledge extracting module and an evidence semantic knowledge extension module are established between evidence receiving and evidence output, wherein data interaction is carried out between the target semantic database and the evidence receiving module, the evidence semantic knowledge extracting module and the evidence semantic knowledge extension module, the evidence receiving module receives object identification evidences from sensor identification sources of different types in real time, an evidence semantic knowledge cluster module carries out clustering on elements in multiple attribute collection related to an expanded attribute constraint relation to obtain a plurality of attributive classifications, an evidence compounding and updating module carries out orthogonal calculation and orthogonal compounding and updating on identification evidences and all attributive classification evidences of different levels of attributes from restricted relationship among multiple attributes to obtain judgment results of updated identification evidences.

The invention belonging to the field of rock exploration geophysics discloses a physical elastic template of a fracture-porosity type rock. In order to study the compact sandstone reservoir with high saturated gas, micro-fracture development and high heterogeneity, the model uses a Voigt-Reuss-Hill model to calculate the elastic modulus of mixed minerals; a differential equivalent medium (DEM) model is used for describing a skeleton elastic modulus of a facture and porosity rock; and on the basis of a Biot-Rayleigh theory, local fluid characteristics of the pore are described and a relationship between the elastic parameters and physical properties of the compact sandstone reservoir is disclosed.

The invention discloses a steady-state characteristic analysis method for a linear induction motor. The steady-state characteristic analysis method includes the following steps: 1), creating an unidimensional magnetic-density distribution model of the linear induction motor; 2), calculating the magnetic density of a y-axis according to the unidimensional magnetic-density distribution model; 3), acquiring electric field intensity through the magnetic density of the y-axis according to characteristics of magnetic barrier surfaces; 4), calculating plural forms of amplitudes of a normal magnetic-density travelling wave, an inlet-end magnetic-density reflecting wave and an outlet-end magnetic-density reflecting wave; 5), calculating impedance of the normal magnetic-density travelling wave, the inlet-end magnetic-density reflecting wave and the outlet-end magnetic-density reflecting wave; 6), calculating total impedance of a single-phase circuit of the linear induction motor; 7), analyzing performance parameters of the linear induction motor in different stable states. By the steady-state characteristic analysis method, analysis difficulty and calculation time of the magnetic characteristics of the linear induction motor are reduced; the steady-state characteristic analysis method is high in accuracy and can be conveniently applied to initial electromagnetism optimization design and middle-late-period electromagnetism parameter and driving performance analysis of the linear induction motor.

The invention discloses a bathtub failure curve based hydroelectric unit deterioration evaluation method. The method takes full account of a bathtub failure path of a unit from installation to final extended service. Three different deterioration models are adopted to evaluate single indicators, so that early, middle and late deterioration degrees of the unit can be accurately characterized by different indicators of the unit. Meanwhile, family defects, inspection records and other site data are considered, and variable-weight weighted calculation is performed on indicators of various aspects, so that the comprehensive deterioration degree and grade of the unit are finally obtained, and maintenance decision and extended service running of the unit is guided. The method can effectively evaluate the deterioration degree of the hydroelectric unit, and has very good economic benefits and engineering values.

The invention provides a method for quantitatively analyzing underground water numerical simulation uncertainty based on information entropy. The method comprises the steps that the information entropy of predictive variable probability distribution is adopted as the uncertainty of a variable, and according to the predictor formula and information entropy theory of a Bayesian model averaging method, the underground prediction uncertainty is decomposed into a model structure, model parameters and the overlapped uncertainty among various concept model prediction distributions. The uncertainty of each probability distribution type random variable can be measured, the defect that a traditional variance method can only measure normal distribution is overcome, and the application range of quantitative uncertainty analysis is enlarged; the underground water numerical simulation uncertainty breaks up into model parameters, the model structure and the overlapped uncertainty, and the defect that the overlapped uncertainty cannot be described through the traditional variance method can be overcome; the model parameter uncertainty is defined as the difference obtained by subtracting the model overlapped uncertainty from the sum of various concept model interior (parameter) uncertainty weights, and therefore the uncertainties are described more accurately and reasonably.

The invention discloses a syntactic information fused quantum language model modeling method, which starts from a quantum system for representing sentences, and introduces syntactic structure information into a model so as to improve the modeling capability of a quantum language model. Comprising a data preprocessing and syntactic structure generation module, a syntactic-guided word representation module and a semantic and syntactic information fused sentence representation module. The data preprocessing and syntactic structure generation module processes sentences into a syntactic structure tree form; the syntax-guided word vector expression module is used for training word vector expression of words by utilizing dependency information of each sentence; a semantic and grammar information fused sentence expression module converts the word vector generated in the previous step into a density matrix, and sentence expression guided by syntax is further obtained and used for training of downstream tasks; on the basis of generation and expression of a syntactic structure tree reconstruction density matrix, lexical term dependency of sentences is modeled by utilizing dependency relationships.

The invention provides a dynamic QoS selection method and system for user demands, and the method comprises the steps: dividing evaluation indexes into hard demand indexes and non-hard demand indexes,carrying out the preliminary screening of services according to the hard demand indexes, and determining a candidate service set; converting the QoS attributes of the candidate services and the QoS demands of the user into a QoS attribute interval of the candidate services and a QoS demand interval of the user, and performing relative dominance calculation; performing candidate service reductionbased on a Skyline method, and determining a Skyline service set; and performing service sorting on the Skyline service set based on a TOPSIS method, and determining an optimal service. According to the invention, the reliability and efficiency of service selection are improved.

The invention relates to a deterioration law assessment technology of a primary power distribution device, in particular to a Markov chain model-based method for describing a state deterioration process of a power distribution transformer. The method comprises the following steps of 1, collecting and sorting deterioration historical records of devices same in type; 2, determining a state space which describes a deterioration degree; 3, establishing a statistical sample; and 4, calculating a state transfer probability. According to the method, a state parameter of the deterioration degree of the power distribution transformer can be definitely expressed; the reliability of expressing the deterioration degree by the state parameter is supported by a fault mechanism; and in addition, a brand-new state and a function fault state of a system can be defined by using a state parameter value which can be observed, so that the description of the state deterioration process of the power distribution transformer is objective, reasonable and accurate.

The invention discloses a network security index calculation method. The method comprises: quantifying network events that qualitatively describe network states, defining the network events as abnormal events and secure events, determining the number of the abnormal events and / or secure events in a network event system and the risk grade of each event by means of an event set, calculating an event grade index, finally calculating a network security index according to an event scale index and the event grade index so as to obtain a description of explaining system stability. The method can converts describe network qualitative description into quantitative description so as to reflect the basic fact of the network security state.

The invention discloses a reliability analytical method of an extra-high voltage direct-current transmission project thyristor converter valve system. The reliability analytical method comprises the steps that 1 reliability calculation of a single valve of the extra-high voltage direct-current transmission project thyristor converter valve system is carried out; 2 the reliability calculation of 12-pulse converter is carried out; 3 the reliability calculation of single-pole converter is carried out; and 4 the reliability calculation of double-pole converter is carried out, and an index capableof assessing the reliability of the extra-high voltage direct-current transmission project thyristor converter valve system is obtained. According to the reliability analytical method of the extra-high voltage direct-current transmission project thyristor converter valve system, a multilayered structure and a function range of a converter valve control system are taken into full consideration in the analysis process, and influences of each level on the operation and reliability of the converter valve are fully reflected.

The invention relates to a quality assessment and quality restoration method for typhoon data. The method comprises the following steps: step S1, obtaining original typhoon data; step S2, analyzing characteristics of the typhoon data and describing the typhoon data; step S3, assessing quality of the typhoon data, wherein the step S3 includes the step S31: assessing the integrity of the typhoon data; the step S32: assessing the consistency of the typhoon data; step S4: restoring the typhoon data according to the sequence of firstly restoring integrity and then consistency of the data, wherein the step S4 includes the step S41: restoring the integrity of the typhoon data; the step S42: restoring the consistency of the typhoon data. The quality assessment and restoration method for typhoon data has the following advantages: quality of the typhoon data with multiple essential factors can be judged and restored, which lays a foundation for similarity measurement of typhoon time; therefore,similar typhoons are more accurate to judge; the method helps make quick decisions for departments concerned in order to make precautions; and as a result, economy loss and deaths and injuries of people are reduced.

The invention discloses an optimization method for angle measurement eccentric errors of collimators. The optimization method includes a parameter setting step: setting parameters including eccentricerrors, measurement angles, actual rotation angles, radiuses and the like; a relation building step: building logic relations between the various parameters; a relation optimizing step: optimizing thelogic relations between the various parameters; a model generating step: utilizing the optimized logic relations between the various parameters as error correction models; an error optimizing step: measuring angles by the aid of the error correction models to optimize the angle measurement eccentric errors. The optimization method has the technical advantages that the mathematical models are built for the eccentric errors of optical grating code disks of the collimators, system defects can be overcome by the aid of digital optimization processing modes, accordingly, influence of the eccentricerrors on the instrument precision and performance can be reduced, and the measurement precision of the collimators can be enhanced; the manufacturing and adjusting difficulty of the optical gratingcode disks can be lowered, the production cycles can be shortened, and the like.

A method for operating a navigation system (11) of a motor vehicle (10), wherein reception of a destination data record (1) containing a geodetic position (2) and a first address (5) associated with the position (2) is followed by geodetically associable digital map data (13) that are available to the navigation system (11) being taken into account in order to determine test addresses, containing at least one road name, that can be reached with the motor vehicle (10) in a destination region (9) around the position (2), and a second address, which differs least from the first address (5) in terms of at least one character string describing the test address, being selected from the test addresses, wherein when the level of discrepancy in the second address from the first address (5) is below a maximum value or corresponds thereto, the navigation system (11) is operated for navigation to the second address.

The invention provides a modeling method for a motor-driven irregular expansion object of an unknown turning rate. The method comprises the steps of 1, using Minkowski addition to decompose the complex motor-driven evolution process of the irregular expansion object H into the combination H1+H2+...+Hn of simple motor-driven evolution processes of a plurality of sub-objects H1, H2, ..., Hn; 2, when the turning rate is unknown during the motor-driven process, establishing the general evolution model for each sub-object by jointly using the turning rate, the extension form parameter and the motion state parameter; and 3, using the general evolution model and the measurement equation to build a common system model. A tracking algorithm for the motor-driven irregular expansion object of the unknown turning rate is applied to the general system model. The invention can not only solve the description problem that the irregular expansion object movement state evolutes over time in a scene that the turning rate is unknown in the typical motor-driven process, but also can fully characterize the abrupt change of the extension form, so that the whole motor-driven process can be reasonably described.

The embodiments of the invention disclose a description method based on a regional image feature concerning network and nearest neighbor ordering. By combining global and regional image feature information, the image description can output overall information and highlight detail information; a double-layer semantic layer for mining deep semantic information and a stack type portal circulation unit for increasing the vertical depth of a recurrent neural network are designed to learn deeper semantic mapping between images and words; and the generated candidate descriptions are reordered by adopting a nearest neighbor algorithm and a semantic similarity, so that the finally output sentence description is more reasonable.

The invention discloses a concentrator model considering energy quality, which comprises the following steps of: based on an energy concentrator model, analyzing energy forms of input and output ports of an energy station, and establishing a coupling matrix; the effective energy coupling characteristic of the energy station is mathematically represented; establishing a hub model based on the coupling matrix and the input and output port column vectors of the energy station; the relation between the input port and the output port of the energy station can be represented, and the input port can be solved based on the output port and the line concentration model; a graph theory is introduced on the basis of a hub model, equivalent nodes and branches are established, the relation of internal distribution of an energy station is analyzed, and a standardized hub model is established and can be used for optimizing the internal distribution of the energy station and the energy quality; and introducing a distribution matrix on the basis of the standardized concentrator model, providing an energy station distribution calculation method based on the standardized concentrator model, and solving the internal distribution of the energy stations.

The invention discloses a fault fracture zone modeling method. The method includes the following steps: step one, identifying the internal structure of a fault fracture zone; step two, using two surface models to represent fracture zone boundaries of both sides of the fault fracture zone, the distance between the two surface models being determined by the width of the fault fracture zone; step three, applying a corresponding grid processing method to the fault fracture zone according to positional relations between the fault fracture zone and an ascending wall and descending wall and the internal structure of the fault fracture zone; and step four, performing gridding again to obtain a volume model of the fault fracture zone. The fault fracture zone modeling method in the invention can accurately simulate the internal structure of the fault fracture zone and blocking of different parts, establish a fault model that better satisfies an actual geological condition, and thus can be widely applied to the field of reservoir description.