153 results about "Artificial neural network algorithm" patented technology

Disclosed is a bridge construction and maintenance whole-process intelligent monitoring, assessment, alarming and decision-making system and method. The bridge construction and maintenance whole-process intelligent monitoring, assessment, alarming and decision-making system is composed of an intelligent collecting module and an intelligent assessment, alarming and decision-making module, wherein the intelligent collecting module integrates laser radar three-dimensional high-precision scanning technology, building information model technology and wireless intelligent sensing technology and intelligently collects multi-source heterogeneous multi-dimensional high-precision information during a bridge construction and maintenance whole process such as environment monitoring information, structure monitoring information and project process quality safety management information; the intelligent assessment, alarming and decision-making module integrates big data processing technology, cloud computing technology, artificial neural network algorithms and virtual reality technology and structures multi-stage safety alarming indexes and emergency decision-making database to perform online, real-time, high-precision, visual and intelligent assessment analysis, alarming and decision-making on structure and traffic safety performance of the bridge construction and maintenance whole process.The bridge construction and maintenance whole-process intelligent monitoring, assessment, alarming and decision-making system covers the bridge construction and maintenance whole process and has the advantages of being high in efficiency and precision, visualized, intelligent and the like.

The invention discloses a BP and GA based blade machining cutting quantity optimization selection method. The method comprises: establishing an aluminum alloy blade milling model by taking a spiral milling process of a ball end mill as a research object; simulating milling of blades of an aeroengine by using ABAQUS nonlinear finite element analysis software; summing up a change pattern between a cutting parameter and a deformation amount in milling simulation and a relationship between the cutting parameter and the deformation amount by using an artificial neural network algorithm, and establishing a blade milling deformation prediction model; and determining a design variable, a target function and a constraint condition for cutting parameter optimization, constructing a cutting parameter optimization model of aluminum alloy blades, and optimizing the cutting parameter by using a genetic algorithm. According to the method, the functional relationship between the deformation amount and the cutting parameter is established, and the cutting parameter optimization model is established, thereby providing a reasonable and quick cutting parameter selection method to guide actual production.

Disclosed is an apparatus for measuring movements of a user, which may include: a sample pattern storing unit for storing one or more sample patterns which are obtained by patterning EEG waveforms and EMG values; a communication unit for receiving an EEG waveform and an EMG value of the user measured by the EEG and EMG sensors; a pattern normalizing unit for selecting sample patterns with the highest similarity to the received user's patterns among the stored sample patterns, and for normalizing the received user's patterns so that their amplitudes and periods coincide with those of the selected sample patterns; a movement estimating unit for estimating movement of the user by inputting the normalized EEG waveform and EMG value in an artificial neural network algorithm; and a movement determining unit for calculating the user's movement by applying each weight value of the EEG and EMG to each movement estimation value.

The invention discloses a two-dimensional laser galvanometer scanning system based on binocular stereo vision, comprising a laser galvanometer system and a calibration device. The calibration device is composed of a linear displacement system and a binocular stereo vision system. The invention further discloses a calibration method of the two-dimensional laser galvanometer scanning system based on binocular stereo vision. The laser galvanometer system is calibrated by building a mapping relationship between specific input digital signals and the direction and position vectors of laser beams corresponding to the signals under a binocular system coordinate system with the aid of an artificial neural network algorithm. The calibration method needs no complex galvanometer system physical parameter modeling process. The calibration result is highly precise and reliable. Moreover, the calibration method can be used in various practical applications based on a galvanometer system, such as laser positioning projection and object surface three-dimensional shape measurement.

The embodiment of the invention provides a volume adjusting method and a mobile terminal. The method comprises the following steps of: calculating an environmental quantitative value of the noise level of an environment, where the mobile terminal is; calculating an application quantitative value of an application program for playing audio in the mobile terminal; calculating an earphone state quantitative value for representing the connection state between the mobile terminal and an earphone; on the basis of the environmental quantitative value, the application quantitative value and the earphone state quantitative value, calculating to obtain a first target volume through a model constructed by an artificial neural network algorithm; and adjusting the play volume of the application program for playing the audio to the first target volume. According to the embodiment of the invention, the output volume can be matched with the noise environment, the application program for playing the audio and the condition whether the earphone is worn or not; therefore, the use experience of users is improved; and thus, the mobile terminal can flexibly adjust the output volume while switching different noise environments, playing the audio by using different application programs or using the earphone or publically playing the audio.

The invention discloses a positioning method of a small unmanned aerial vehicle based on binocular vision, which includes the following steps: using a ground target as a reference, selecting the target centroid as the origin of the navigation coordinate system, wherein the binocular camera fixed on the computer is based on the V4L2 interface to collect the image of the object in real time; By means of the object detection algorithm based on color model and artificial neural network algorithm to remove the disturbance region in the image, extracting the target area from the left and right viewsprecisely, the difference in the width of the target centroid pixel between the two images is the parallax of the target point; combined with binocular ranging and camera calibration parameters, theposition of the target point in the left camera coordinate system is calculated, and then the rotation matrix from the current airframe coordinate system to the initial airframe coordinate system is calculated based on the attitude angle information obtained by IMU, so as to calculate the coordinates of the UAV in the navigation coordinate system. The method effectively shortens the positioning time of the unmanned aerial vehicle (UAV) and is beneficial to the real-time processing of the aircraft position.

The invention discloses a photovoltaic cell non-destructive detection method and a photovoltaic cell non-destructive detection system based on induction photothermal radiation. According to the present invention, electromagnetic induction excitation is performed with an excitation coil, and the electricity, the magnetism, the light, the heat and other information of the photovoltaic cell and the assembly are subjected to high-speed and precision measurement in a completely non-contact manner; the attribute, the defect and the health degree of the photovoltaic cell and the assembly are subjected to subtle and quantitative evaluation by fusing machine learning and an artificial neural network algorithm, such that the disadvantages that the subtle defect cannot be detected, the in-service detection cannot be achieved, the detection is slow and the like in the existing contact type detection are solved; the performance, the defect type and the whole health degree of the photovoltaic cell are subjected to quantitative evaluation by comprehensively utilizing the multi-frequency impedance information, the optical radiation information and the thermal radiation information so as to provide the theory, the method and the technical support for the ordered operation of the photovoltaic cell industry chain; and the method can be used for the online diagnosis of the failure of the photovoltaic cell and the assembly installed in the photovoltaic power plant so as to substantially improve the detection efficiency and the safety of the photovoltaic cell and the assembly.

The present invention relates to a coagulation dispensing control method and system based on an artificial neural network algorithm. The method comprises the following steps: S1, building a nerve network; S2, performing coagulation dispensing sample obtaining and pretreatment, and obtaining a sample value; S3, initializing a nerve network weight, and inputting the sample value at a nerve network model; S4, calculating an input layer, a hidden layer, an output layer and a carrying layer number value; S5, calculating the function error of the nerve network and updating the nerve network weight, performing training of the nerve network, and completing the study of the nerve network; and S6, determining whether the updated weight satisfies the setting precision or the number of times of training or not, and performing the predication of the real quantity of reagent. The automation of the coagulation dispensing technology and the online monitoring of the production operation parameters are realized, and the coagulation dispensing control method and system based on the artificial neural network algorithm provides the guarantee for the safe production of a water plant and reach the purpose of saving the medicine consumption, decreasing the manpower and reducing the labor intensity of operators.

The invention discloses a system based on a TCM syndrome differentiation artificial neural network algorithm model, which can enhance the syndrome differentiation classification capability of different dominant diseases, improve the efficiency and accuracy of TCM syndrome differentiation diagnosis and treatment, can expand the scope of treated diseases of TCM doctors, improve the diagnosis and treatment capability of incurable diseases, and shorten the clinical experience accumulation time of young TCM doctors. The system includes a symptom input layer module for receiving the input patient disease data; a first hidden layer module for performing quantization coding of the input disease data; a second hidden layer module for performing correlation function calculation according to the quantization coding of the disease data to obtain the corresponding disease cause, disease location, disease nature and disease characteristic classification; and a syndrome-type output layer module for outputting the result data of the symptom and the corresponding syndrome type when the coincidence of the disease cause, disease location, disease nature and disease characteristic of the disease and the nature and flavor characteristic of the medicinal material is higher than the preset threshold.

The invention discloses a power business management method and system based on a 5G network slice. The method comprises the following steps: collecting corresponding data information under different business scenes by using a power terminal; transmitting the data information to each power business platform based on a network slice of an operator, and preprocessing the data information; and constructing a data analysis model by using an artificial neural network algorithm, and analyzing and managing the preprocessed data information based on the model to finish the management of the power business. According to the invention, the requirements of an intelligent network on the aspects of service isolation, ultralow time delay, ultrahigh reliability, safety and the like of a communication network are met, slices with different scene requirements in the power industry can be applied, ordered, configured, monitored, operated and maintained through the 5G network slice management system in the power industry, and network slices of different operators can be managed; network-as-a-service is realized, and customized network communication requirements of the power industry are satisfied.

The invention discloses a distribution transformer load forecasting method based on a mxnet framework deep neural network, and relates to a distribution transformer load forecasting method. At present, there is no uniform method to characterize and describe the distribution transformer and a worker cannot fully understand the distribution transformer and cannot accurately forecast the load changetrend. The method includes the following steps: acquiring system internal data and external data; purifying the acquired data to obtain load-related index data and historical load data; and separatelyfitting a long-term load forecasting model and a short-term load forecasting model in the load by using an optimal combination prediction model and an artificial neural network algorithm and using acourt as a unit; according to a load forecasting result and an index dimension, extracting a load-related label system and building a view of the court; displaying portrait view on a human-machine interface. The method establishes a load forecasting assessment model to monitor load fluctuations, achieve continuity of forecasting, grasp the important characteristic-load dynamic change process of acommon transformer, and can forecast the load change trend.

The invention discloses a method and a device for digitally detecting the quality of edible vinegar. The device comprises a sample container which is placed on the upper surface of a conveying belt, can move forwards along with the conveying belt and contains an edible vinegar sample, a vinegar color and state information acquisition component, a vinegar aroma information acquisition component and a vinegar taste information acquisition component are sequentially disposed at suitable positions, the sample container moves forwards and passes through the suitable positions, the vinegar color and state information acquisition component, the vinegar aroma information acquisition component and the vinegar taste information acquisition component respectively acquire color and state information, aroma information and taste information, the three types of information are fused to form a data matrix, a principle component factor is extracted from the data matrix to be used as an input vector of a judging model by means of principle component analysis, a multi-variable progression model between the principle component factor and edible vinegar sample grades evaluated by professionals is built by an artificial neural network, and digital information of the flavor and the quality of the edible vinegar to be tested can be obtained according to the built model. The method and the device have the advantages that objective, accurate and quantitative representation of the flavor and the quality of the edible vinegar is guaranteed, and integrity and completeness of the representation of the flavor of the edible vinegar are ensured.

The invention discloses a refrigerant leakage detection method and a refrigerant leakage detection device for an air conditioner. The refrigerant leakage detection method comprises the steps of: acquiring current operating parameters of the air conditioner and environment information of an environment in which the air conditioner is located; inputting the current operating parameters and the environment information into a trained neural network model to obtain a refrigerant residual amount output by the neural network model; and determining whether the air conditioner has refrigerant leakage according to the refrigerant residual amount. By adopting the refrigerant leakage detection method and the refrigerant leakage detection device, the effects of detecting air conditioner refrigerant leakage and improving the accuracy by adopting an artificial neural network algorithm are achieved.

The invention discloses environment-friendly type oil for a heavy-duty industrial gear. The oil comprises base oil, an anti-wear reagent at extreme pressure, an antioxidant corrosion-resistant agent, a metal passivator, an antirust agent, a demulsifier, an antifoaming agent, an oiliness agent, and a viscosity index improver. The oil for the heavy-duty industrial gear is good in comprehensive property, not only has good antioxidant and corrosion-resistant properties, antirust property, anti-wear and corrosion-resistant properties, friction reduction, bearing, high torque, high-temperature stability, demulsibility, antifoaming property and the like, and is capable of reducing emission, good in biodegradability, and lower in cost, but also can be used for enabling an industrial gear to form an anti-wear, high temperature-resistant and extrusion-resistant boundary lubricating oil film under the high torque or impact load, so as to ensure that the gear is well lubricated. Additives are selected by adopting an artificial neural network algorithm, the optimal formula can be selected by a genetic algorithm, and the environment-friendly type oil for the heavy-duty industrial gear can meet the performance requirements of the GB5903-1995 heavy-duty industrial gear on the oil with the viscosity grades of 100, 150, 220, 320, 460 and 680.

The invention discloses a network security authentication system, comprising a data capturing module, used for capturing suspicious traffic and abnormal behaviors of an import and export system to form sample data; a data preprocessing module, used for carrying out screening preprocessing on the sample data, removing the noise data in the sample data and carrying out dimensionality reduction processing on the sample data; and a behavior analysis module, used for carrying out clustering analysis on the preprocessed sample data by adopting an improved K-Means clustering method, and carrying out aggressive behavior detection based on an artificial neural network algorithm to identify potential unknown attacks in the network. The network security authentication system disclosed by the invention can detect known and unknown network attacks, carry out clustering analysis on the preprocessed suspicious traffic and the abnormal behaviors of the import and export system by adopting the improved K-Means clustering method and can accurately separate the network attacks of various types, so as to achieve very high accuracy and a very low false alarm rate.

The invention provides a method for identifying urban power grid feeder load based on an artificial neural network. The method comprises the following steps: (1) searching equipment composition, historical power consumption information and meteorological information of typical construction load of a corresponding climatic region; (2) establishing power consumption load models of all kinds of typical constructions; (3) generating load curves of corresponding time periods of all kinds of typical construction loads based on date and meteorological information of a to-be-identified feeder load curve; and (4) identifying the load composition of the to-be-identified feeder load curve by using a BP artificial neural network algorithm. According to the method provided by the invention, different construction load composition of large feeder nodes is effectively identified, which is conducive to predicting the dynamic load properties of the feeder nodes and evaluating the demand response potentials, and foundation is laid for regulating a reasonable demand response policy and mechanism.

The invention discloses a device and a method for detecting water content of sand-gravel aggregate. The device comprises a water content sensor data analysis module, a central processor module, an artificial neural network processor module, a water content data display and management module and a water content signal transformer module which are successively connected with each other, wherein thewater content sensor data analysis module is connected with a plurality of sensors; the sensors comprise two water content microwave sensors, two water content capacitance sensors and a temperature sensor; and the sensors are connected with the central processor module through the water content sensor data analysis module. The method comprises the following steps: simultaneously measuring parameters such as water content signals and material temperature of a plurality of material positions of the sand-gravel aggregate in a sand-gravel aggregate bin by using the water content sensor data analysis module, then calculating the water content of the sand-gravel aggregate by using a data fusion technology based on an artificial neural network algorithm so as to achieve seal-time detection of thewater content of the sand-gravel aggregate in a concrete batching process and improve stability and reliability of water content measurement data.

IOP sensors and IOP measurement algorithms are disclosed herein. Examples of the IOP measurement algorithm can include a signal demodulation and artificial neural network algorithm to produce reliable results using minimal computational resources. The overall accuracy and speed of the IOP sensors and measurement algorithms enable their implementation in home-based IOP measurement systems.

The invention provides a garbage sorter based on speech recognition and a working method thereof. The garbage sorter comprises a garbage can main body, a garbage containing box, a main control unit, aspeech recognition unit, an audio unit, a storage unit, a network unit and a sensor unit, wherein after detecting that someone is close to the garbage sorter, the system enters a working state, collects speech data in real time, uses artificial neural network algorithm to process and recognize the speech data, outputs the corresponding speech information, and uses BP neural network algorithm to obtain the garbage type according to the speech information, the audio unit plays the name of the corresponding garbage type, and lights an indicator light, opens the opening of the garbage containingbox, and finally sends the usage of the garbage sorter to a cloud platform for recording. According to the invention, the speech recognition technology and the BP neural network algorithm are added tothe process of garbage sorting to realize intelligent management of garbage sorting, improve the efficiency of garbage sorting, and help people understand garbage sorting more accurately.

The invention belongs to the technical field of design of aircrafts and in particular relates to an artificial-neural-network-based inverse design method for aircraft airfoils / wings. The method comprises the following steps: rebuilding expression modes of airfoils / wings by using PARSEC parameterization method of airfoils / wings, and implementing inverse design technology by using an artificial neural network (ANN) algorithm. By virtue of the artificial-neural-network-based inverse design method for aircraft airfoils / wings, a conventional cut-and-try method with complex design and low efficiency for aircraft airfoils / wings is put aside; a relation between the aerodynamic performance of airfoils / wings and the geometric profile of aircraft airfoils / wings is directly built; the artificial-neural-network-based parameterization inverse design for aircraft airfoils / wings is implemented. The artificial-neural-network-based inverse design method for aircraft airfoils / wings is high in speed and is highly suitable for the overall design of aircrafts, especially the initial design; the artificial neural network algorithm is in place, so that the generated results are highly accurate.

The present invention discloses a material quality assessment method based on machine learning. A material quality assessment model is created, the artificial neural network algorithm of back propagation is selected, and a back propagation network model is established. The collected historical sample information is divided into two portions consisting of training data and verification data. The training data is used to perform training of the model to obtain model weight parameters and regulation factors. The verification data is used to perform test verification of a model training result until a verification result accords with collected information so as to obtain the non-linear relation of material quality and influence factor quality testing. The verified model and the model parameters are employed to predict the quality level of the materials. The material quality assessment method based on machine learning can more objectively and accurately assess the quality level of the materials.

The invention relates to a median sagittal plane construction method based on an artificial neural network model. The method comprises the following steps: (1) establishing an artificial neural network model suitable for constructing the median sagittal plane through a three-dimensional face data PA association method; and (2) establishing an intelligent weight regulation and control mechanism ofthe artificial network model aiming at the PA anatomical mark point, and realizing a training mode and a feedback mechanism of the expert-level logic strategy. According to the method, the artificialneural network algorithm model can be applied to symmetry analysis of oral medical three-dimensional face data, an algorithm idea of constructing a median sagittal plane through an 'artificial neuralnetwork PA association method' is innovatively provided, intelligent weight regulation and control of PA mark points are achieved, and the effect of simulating a clinical expert-level diagnosis strategy is achieved.

The invention discloses an optical fiber eavesdropping positioning method and system based on machine learning and electronic equipment, and the method comprises the steps: collecting the historical signal information of an optical fiber channel, and the historical signal information comprising eavesdropping condition information and channel performance information; training by using an artificialneural network algorithm according to the historical signal information to obtain an artificial neural network model for eavesdropping positioning; collecting current signal information of the optical fiber channel, and judging whether the current signal information has an eavesdropping behavior or not; and if the eavesdropping behavior exists, taking the current signal information as the input vector of the artificial neural network model, and determining the position of the eavesdropping point according to the output vector of the artificial neural network model. According to the invention,refined processing is carried out on optical communication physical layer data by using an artificial neural network algorithm, thereby realizing intelligent positioning of optical fiber eavesdropping; and comprehensive analysis and calculation are carried out through the eye pattern of the optical fiber channel and the parameter difference, and eavesdropping positioning is carried out.

The invention provides a method for identifying Pingli fiveleaf gynostemma herb through near infrared spectroscopy, which comprises the following steps: (A) establishing a near infrared spectroscopy identifying model of Pingli fiveleaf gynostemma herb: (A-1) selecting the spectral range of 4,000-12,500cm<-1>, and scanning the near infrared spectrogram of the Pingli fiveleaf gynostemma herb; (A-2) pretreating the data of the spectral range of 4,000-9,500cm<-1>; (A-3) extracting the main components; and (A-4) establishing an artificial neutral network model: determining the structure of the neutral network by an artificial neutral network algorithm according to the characteristics of the input and output data, and training the neutral network by use of the training data; and establishing a BP artificial neutral network model of input layer nodes 10-, hidden layer nodes 5- and output layer nodes 2 by use of the MATLAB software; and (B) identifying the unknown sample: scanning the near infrared spectrogram of the unknown sample under the same conditions, selecting the number of main components, judging the authenticity of the unknown sample according to the trained neutral network model, and representing the output nodes with binary codes respectively, wherein 10 represents Pingli fiveleaf gynostemma herb, and 01 represents non-Pingli fiveleaf gynostemma herb.

The invention discloses an SQL loophole detecting method based on artificial neuron network. The method consists of an SQL statement analysis module and an artificial neuron network detection module. The main function of the SQL statement analysis module is to convert the SQL statement submitted by the Web application into a feature vector group r composed of 0 and 1, and then submit the resulting feature vector to the artificial neuron network detection module for analysis and detection, and fully apply the self-learning function of the neuron network to realize the detection of SQL statement through the artificial neuron network algorithm obtained by training and regex match. Compared with the traditional method, the SQL loophole detecting method is more convenient, accurate and fast, and does not need to modify the code itself. In addition, the system is developed by Javan language only, which is easy to maintain and can be used for secondary development. The system can well realize function expansion and cross-platform migration and has great application value for SQL injection loophole detection and research.

The invention discloses a Delphi software-based method for realizing a neural net algorithm, which comprises the following steps: after establishing an ActiveX automatic connection between a Delphi application program and a Matlab, executing a Matlab command in the Delphi to realize effective database calling by using the Matlab; performing the neural net learning analysis of the effective data; and feeding back an algorithm result to the Delphi in real time and successfully applying the algorithm result to actual control. In the method, the Matlab neural net algorithm is successfully applied to a control program of the Delphi, and the algorithm executing efficiency of the Matlab is improved considerably; and at the same time, the problems of seamless interfaces and unified packing of an artificial neural net algorithm program and other application programs are solved, so the neural net algorithm program has high reliability, unalterability and security.

The invention discloses an artificial neural network algorithm based indoor environment negative feedback adjustment system. The artificial neural network algorithm based indoor environment negative feedback adjustment system comprises a data processing platform, an intelligent control system, an indoor environment monitoring system, a human thermal sensation feedback system and a basic item parameter input system; and the data processing platform is connected with the intelligent control system. according to the artificial neural network algorithm based indoor environment negative feedback adjustment system provided by the invention, each weight matrix is obtained after environment parameters and human hot sensation feedback values are subjected to feedback calculation through a model, and a scheme of improving the human comfort evaluation indexes is provided after the intelligent control system processes the weight matrixes, and correction values of indoor environment parameters areobtained through data processing; the artificial neural network algorithm based indoor environment negative feedback adjustment system can monitor the indoor environment wholly, can realize feedback automatic adjustment, has a learning function and can obtain the optimal indoor environment parameters automatically based on different environments.

The invention discloses a radiation temperature inversion method based on an artificial neural network algorithm. The method specifically comprises the following steps: constructing an artificial neural network model; simulating and generating a sample data set by utilizing a one-dimensional laser target coupled fluid program; preprocessing the sample data set; training an artificial neural network model by using the preprocessed sample data to obtain an optimal artificial neural network model; and carrying out a shock wave experiment on a sample material to obtain a target back light-emittingimage of the sample material having been impacted, and acquiring radiation unloading temperature evolution of the sample material by utilizing the optimal artificial neural network model according tothe target back light-emitting image. According to the method, inversion of the radiation temperature under metal impact loading is realized by utilizing the artificial neural network, and the methodis a novel method for calculating the metal impact radiation temperature under impact loading and has important significance for theoretical modeling of a high-pressure state equation.

A precision calibration method of a load moment limiter which timely measures and displays crane working condition parameters as the weight, the amplitude, the length, the angles and the like. The method comprises the following steps: adopting the self-adapting calibration technology based on artificial neural net algorithm; through transforming a circuit, a single chip microcomputer and a memoryto automatically collect and store the working condition parameters in different motion states (upward and downward or stillness) and timely operating and treating data, obtaining precision value of the weight and the amplitude of the crane in the working condition; having no need for operating and treating data additionally; effectively excluding the errors due to man-made interference; and greatly improving the efficiency and the precision of the load moment limiter. The method can cause the load moment limiter to precisely calibrate according to the practical use condition, improve the precision and the debugging efficiency of the load moment limiter, effectively guarantee the operating safety of the crane, and is widely applicable to the precision calibration of load moment limiters of mobile cranes (an automobile crane, a crawler crane and a tyre crane) and non-mobile cranes.