156results about How to "Realize online identification" patented technology

The invention relates to an image recognition method based on multi-column convolutional neural network fuzzy evaluation. By combining an image preprocessing technology, a convolutional neural network and a fuzzy mathematical method, interference information in an image sample acquired in the real environment is overcome by adopting different preprocessing technologies from different perspectives respectively, feature information beneficial to recognition is strengthened, the preprocessed image is input to the preprocessed and optimized multi-column convolutional neural networks having different structures respectively, then fuzzy evaluation based on comprehensive affiliation degree and discrete affiliation degree is performed on the output results of the multi-column convolutional neural networks by adopting a fuzzy matrix to determine a fuzzy evaluation manner suitable for the image, and a recognition result is finally accurately obtained.

The invention discloses an airborne distribution type POS (position and orientation system) transfer alignment method based on parameter identification. An airborne distribution type POS comprises a high-precision master POS and a plurality of sub IMUs (inertial measurement units). Aiming at the problems of unknown elastic deformation and time change of a machine body, the elastic deformation angle of the machine body is regarded as a second-order Markoff process, the elastic deformation angle and the elastic deformation angle speed are expanded into state variables, the differences between the speeds and the postures of the master POS and the sub IMUs are measured, and a mathematic model of sub IMU transfer alignment is built; then, the Kalman filtering is performed and parameters of the second-order Markoff process are calculated and updated by virtue of the estimated elastic deformation angle and elastic deformation angle speed; at last, the updated second-order Markoff process parameters serve as initial filtering values at next moment to estimate the relatively accurate posture errors, speed errors and position errors of the sub IMUs, strapdown calculation results of the sub IMUs are corrected by virtue of the errors, and the relatively accurate positions, speeds and postures of the sub IMUs are calculated.

The invention discloses an electromobile data acquisition and management system based on a visual instrument. The system is characterized by comprising a vehicle-mounted data acquisition system and an upper computer data management system, wherein the vehicle-mounted data acquisition system comprises a wireless transmission module, a controller area network (CAN) protocol control module, a CAN bus driving module, a high-speed photoelectric coupler, a secure digital (SD) storage module and a global positioning system (GPS) positioning module; the CAN protocol control module is used for acquiring data on a vehicle-mounted CAN bus in real time and packaging the acquired data; the wireless transmission module is used for sending the acquired data to a digital signal processor of the upper computer data management system; and the upper computer data management system comprises a data transceiving module, a data processing module, a driving operating condition automatic generation module and a driving operating condition on-line identification module. By the system, on-line and real-time identification of the driving operation condition of an automobile is realized, reliable reference is provided for real-time adjustment of an electromobile control policy, and the economy, the dynamic property and the smoothness of fuel oil are further improved.

The invention provides a gravity compensation method for mechanical arm load mass and sensor null drift online recognition. The gravity compensation method comprises the steps that S1, a load mass center and mass recognition module is adopted for obtaining and recognizing the mass center and mass of loads at the tail end of a mechanical arm, and the recognized mass center and the recognized mass of the loads are obtained; S2, a sensor null drift recognition module is adopted for reading the numerical value of a sensor, calculation is performed according to the recognized mass center and the recognized mass, and sensor null drift is obtained; and S3, a gravity compensation module is adopted for performing gravity compensation on the numerical value of the sensor according to the sensor null drift. By means of the method, the problem that due to the additional force and torque introduced by the loads of the tail end of the mechanical arm to the tail-end six-dimensional force sensor by the gravity effect and the presence of sensor null drift errors, the numerical value of the sensor is inaccurate and the final force control effect is affected can be solved, and uncertainty of the sensor null drift determines the necessity of online recognition of the sensor null drift.

The invention discloses a method for regulating a propeller direction to point to a combined-body spacecraft centroid, relating to a propeller control method of a combined-body spacecraft in on-orbit service, and the method further comprises a quality and characteristic identification method of the combined-body spacecraft, belonging to the field of posture control of a spacecraft. The method comprises the following specific implementing steps of 1, online identifying a centroid position of the combined-body spacecraft, and obtaining a centroid deviation of a tracking spacecraft and the combined-body spacecraft; 2, obtaining a regulation angle of the propeller according to the centroid deviation; and 3, adjusting the propeller direction according to the regulation angle of the propeller obtained in the step 2, and enabling a spray nozzle of the propeller to repeatedly point to the centroid of the combined-body spacecraft. According to the method, the combined-body centroid can be identified, and the propeller direction can be regulated and is enabled to point to the combined-body centroid, and the regulated propeller can be used for integrated control of a combined-body spacecraft attitude and orbit.

The invention relates to the field of foam flotation automation, and discloses a flotation working condition classification method and a system to provide guidance for subsequent chemical dosing and other operations, and thus improve economic and technical indexes of a flotation process and reduce labour intensity of workers. According to the method, foam images are preprocessed and segmented to obtain all images formed by single bubbles; morphological feature vectors of the images of the single bubbles are extracted; pixel set features of all the image of the bubbles are extracted according to a deep convolutional neural network model, the morphological feature vectors corresponding to the images of the bubbles are combined to finely classify the images of the bubbles, appearance frequency of the various bubbles in the foam images is counted to form a bubble classification frequency set of the single-frame foam images, correspondence relationships between the bubble classification frequency set of the foam images and typical flotation working conditions are analyzed, and thus flotation working condition types reflected by the foam images are obtained.

The solid power switch control method uses MPU and PLD for integrated control and management, and completes the communication with bus of up machine. Wherein, using MPU to integrated control and protect the I2T and PLD to logic synthesize control signals of all channels; using level-control magnetic isolation drive protective circuit with voltage bootstrapping function; sharing one high-frequency demodulation signal for multipath drive; using rating code function to make MPU online recognize rating of every power plate and distinguish different I2T tripping curve. This invention increases power density of solid power switch, and reduces volume and weight.

The invention discloses an on-line rotational inertia identification device for an alternate current permanent magnet servo system and an identification method, belongs to the technical field of servo motor control, and solves the problem that the conventional identification of rotational inertia of all motors is identification of rotational inertia of load mechanical systems in an off state so as not to adjust system control parameters according to current conditions of a system in real time. The device consists of a position controller, a speed controller, a torque controller, a motor and load unit, a photoelectric encoder, a differential unit, a sinusoidal signal generator and a rotational inertia identification unit; and the method comprises the following steps of: injecting a harmonic command with a lower amplitude value to a given position of a position loop motor during production run of a motor system; and extracting a motor position angle and torque control signal of a corresponding frequency component to realize on-line identification of the rotational inertia. The device and the method are suitable for on-line rotational inertia identification of the alternate current permanent magnet servo system.

The invention provides a real-time identification method for a fired coal lower heating value of a coal-fired power plant. The real-time identification method includes the steps of (1) obtaining boiler structure parameters according to a boiler operation design discipline, and reading an operation condition measurement point real-time value at the given moment from a real-time database of a DSC, (2) calculating the specific enthalpy and the density of working media and the specific heat and the density of smoke based on a working medium physical property parameter base and a smoke physical property parameter base, (3) correcting main steam flow masses based on a relation model of main steam adjustment door opening degrees and the main steam flow masses, (4) respectively calculating corresponding energy output of various parts on the boiler side and total energy output at the given moment according to a boiler-side complete-flow-path mechanism model, (5) building a transfer function model between coal feeding flow masses of a coal feeder and the fired coal flow masses, and (6) building a transfer function model between the fired coal flow masses and the total energy output. The real-time identification method is used for on-line identification on the fired coal lower heating value, and has the application potential of combustion and coal blending optimization.

The invention relates to a multi-parameter online identification method of a permanent magnet synchronous motor (PMSM). The multi-parameter online identification method comprises the following steps of: the step 1, establishing a permanent magnet synchronous motor mathematical model under a high-frequency voltage signal; the step 2, calculating the current and dq axis inductance of a PMSM under the excitation of the rotating high-frequency voltage signal; the step 3, calculating the resistance of the PMSM under excitation; and the step 4, calculating the permanent magnet flux linkage of the PMSM. According to the method, the dq axis inductance of the permanent magnet synchronous motor is identified on line by injecting a rotating high-frequency voltage signal, the resistance of the permanent magnet synchronous motor is identified on line by adopting a square wave current signal, and the permanent magnet flux linkage of the permanent magnet synchronous motor is identified on line by adopting a model reference self-adaptive method, so that the multi-parameter online identification of the motor is achieved, the method can be applied to sensorless control, and the control precision isimproved.

The invention discloses a PMU and a method for achieving the on-line identification and alarm of the secondary synchronous oscillation. The PMU comprises an anti-aliasing filter module, an ADC sampling module, a synchronous sampling clock module, a DFT quadrature conversion module, a front-loading filter module, a phasor amplitude analyzing and processing module, and an uploading and processing module. The method comprises the steps of accurately calculating the synchronous phasor; pre-filtering the synchronous phasor; sampling and processing the calculated amplitude of the phasor; analyzing the frequency spectrum of the phasor amplitude; identifying the secondary synchronous oscillation according to the components of the frequency spectrum of the phasor amplitude; uploading the frequency components of the secondary synchronous oscillation and the alarm information to a dispatching master station. According to the technical scheme of the invention, the on-line identification and alarm of the secondary synchronous oscillation of the power system can be completed by the PMU. Through adopting the analysis on the frequency spectrum of measured high-density phasor data, the frequency components of the secondary synchronous oscillation can be accurately calculated. Therefore, the analysis monitoring accuracy and the analysis monitoring adaptability of the secondary synchronous oscillation are improved.

The invention relates to an online recognition method for zip-top can smooth curved surface coding based on machine vision. The online recognition method includes the steps of S1, collecting an original image of zip-top can smooth curved surface coding characters; S2, preprocessing the collected original image; S3, segmenting a coding region in the preprocessed image; S4, segmenting and normalizing the coding characters; and S5, recognizing the characters. The method of the invention can replace the artificial recognition by the machine vision so as to realize the online recognition of the zip-top can smooth curved surface coding with the high precision and high efficiency while ensuring a low detection cost, a low false detection rate and a low detection omission rate.

The invention discloses a firmware version detection method and a vulnerability repair rate evaluation method of an Internet of Things device. The detection method comprises the following steps: S1, acquiring a firmware image file; 2, extracting that file system after decompressing the firmware image files, and determining the Web root directory of each file system; 3, searching Web static resources, respectively extracting eigenvalues of each Web static resource in various firmware versions under each device model, and constructing correspond firmware version characteristic tables; S4, obtaining the IP address list and firmware version characteristic table of the required device type, scanning fingerprint, extracting corresponding fingerprint, and identifying the firmware version number of the device; The evaluation method includes obtaining model version information of the target vulnerability, and calculating the repair rate of the target vulnerability according to the result of thedetection method. The invention has the advantages of simple realization method, high detection accuracy and efficiency, and the vulnerability repair rate evaluation without triggering the vulnerability.

The invention discloses a data drive system state model online distinguishing method based on multi-combination classifiers. The method is used for solving the technical problem that an existing data drive system state model online distinguishing method has modeling difficultly. According to the technical scheme, in combination with analysis of aircraft historical monitoring data and system operating characteristics, through design and training of the multi-combination classifiers, real-time measured data is analyzed and classified, and operating state models of all systems of an aircraft are obtained through distinguishing. The technical problem that due to the high complexity degree of the systems, physical models cannot be directly used for erecting mathematical models is solved. In the distinguishing process, due to the fact that the classifiers are designed and trained in advance, the distinguishing time of real-time models is short, online distinguishing of system state models can be achieved, the models reflect whether the operating aircrafts breaks down or not and the type and degree of the fault, and great significance is achieved for research such as system real-time state monitoring, fault-tolerant control system design and fault recovery.

The invention discloses a transformer area line online identification system comprising a house service wire management terminal which is installed at an electric meter box side and connected with an electric meter in the electric meter box and used for transmitting out meter address identification information of the electric meter and geographical location information of the house service wire management terminal in real time through power line carrier communication; a house service wire transformer area management terminal which is connected with an area transformer and used for transmitting out address identification information and geographical location information of the house service wire transformer area management terminal and the received meter address identification information of the electric meter and the geographical location information of the house service wire management terminal in real time; and a master station system which is used for receiving and obtaining the corresponding relation between the area transformer and the electric meter connected with the area transformer according to the information in real time online so as to realize online identification of the transformer area line. According to the system, the corresponding relation between the area transformer and the electric meter can be timely and accurately obtained online according to the information acquired through power line carrier communication so that later maintenance of the transformer area line and equipment related to the line is facilitated.

The invention provides an online active type cutting chatter restraining system and method for an intelligent milling motorized spindle. The online active type cutting chatter restraining system for the intelligent milling motorized spindle comprises a sensing portion, a controller and a chatter restraining adjustment and control device, wherein the sensing portion, the controller and the chatter restraining adjustment and control device are sequentially connected. The sensing portion comprises a non-contact type eddy current displacement sensor and a signal collector, wherein the non-contact type eddy current displacement sensor is installed at the front end of the spindle and used for measuring the axial displacement of the spindle, and the signal collector is used for collecting displacement signals of the non-contact type eddy current displacement sensor and converting the displacement signals to acceleration signals. The controller is used for outputting voltage control signals to the chatter restraining adjustment and control device according to the acceleration signals output by the signal collector. The chatter restraining adjustment and control device comprises a power amplifier and an active type electromagnetic bearing. The active type electromagnetic bearing comprises a rotor core and a stator core, wherein the rotor core is installed at the front end of the spindle, and a copper coil is wound around the stator core. The power amplifier is used for converting the received voltage signals into current control signals. The output end of the power amplifier is connected with the copper coil. The online active type cutting chatter restraining method for the intelligent milling motorized spindle comprises the steps of identifying and restraining chatter in the machining process.

The invention discloses a multi-fault mode identification method and device of a swashplate of a helicopter. The multi-fault mode identification method comprises the steps that analog quantity signals collected in real time are pre-processed to obtain real-time data corresponding to the analog quantity signals; one or more time domain features in the real-time data are extracted; by comparing each time domain feature with a time domain feather threshold range corresponding to the time domain feature, one or more suspected faults are determined; based on the real-time data corresponding to the one or more suspected faults, one or more fault information items are obtained through a pre-trained radial basis function neural network. The multi-fault mode identification method and device of the swashplate of the helicopter improve real-time performance and efficiency of fault monitoring, eliminate potential fault threats, improve safety of automatic inclination, improve the accuracy of fault identification, shorten the time for fault identification and improve efficiency of fault identification.

The invention discloses a lookup method for specific components of traditional Chinese medicines based on visualized mass spectrometry information. The lookup method comprises the following steps: firstly, carrying out LC-MS analysis on a sample, obtaining quality information of a compound in the sample and retaining information and signal strength; drawing a quality-time diagram of the sample, wherein each measurable compound has a specific spatial position in the diagram; generating an m/z-t-I graph representing the signal strength (I) with the size of the area, which is suitable for computer recognition, and extracting spatial information such as contour information from the graph so as to obtain sample features and compare different samples; establishing a method to obtain repeated chemical component clusters in each traditional Chinese medicine; deducting overlapping components of the chemical component clusters in the two compared medicinal materials and obtaining specific component clusters of each traditional Chinese medicine, which are different from those of the contrast traditional Chinese medicine; carrying out secondary mass spectrometry fragmentation by using the specific components as target compounds and identifying the specific components by database matching.

Provided is a cross-district interconnecting large power network homology identification method based on projection pursuit. Rotate speed information of a generator is read from a wide area measurement system to generate a standardized rotate speed sample matrix, a mathematic model of electrical power system unit homology grouping is established according to the rotate speed sample matrix, the mathematic model of the electrical power system unit homology grouping is solved by adopting a genetic algorithm to determine all projection directions and corresponding target evaluation function values, the optimum projection direction, and the projection contribution rate and the accumulated contribution rate of the rotate speed sample matrix are determined according to projection evaluation function values to determine the homology grouping results of an electrical power system. According to the cross-district interconnecting large power network homology identification method based on the projection pursuit, calculation is simple and easy to achieve, due to the fact that the unit homology grouping is carried out based on the wide range actual measured data of the electrical power system, the influences on the electrical power system unit grouping results by element model parameters and system model parameters in the electrical power system are avoided, and on-line identification of the electrical power system homology unit group is achieved.

The invention discloses an improved particle filter algorithm-based binary state of charge estimation method. According to the method, a traditional Thevenin model is combined with a KiBaM model capable of correctly reflecting rate capacity characteristics of storage batteries to serve as a binary state of charge comprehensive model. By adoption of a recursive least-squares method, model parameters are online identified, and an improved particle filter algorithm is applied to the comprehensive model so as to realize binary state of charge estimation; on the basis of a standard particle filteralgorithm, related steps of a residual resampling algorithm and a Thompson-Taylor algorithm are further imported, and the resampling algorithm is capable of effectively remitting the particle degeneracy problem that sequential significance sampling method must face. Conventional sampling is easy to intensify particle depletion, and new particles are generated through the Thompson-Taylor algorithmto ensure the diversity of the particles, so that the filter performance of systems is strengthened, nonlinear working characteristics of storage batteries can be better adapted, and then correct andreal-time estimation of state of charge is realized.

The invention relates to a traffic sign recognition method based on shape feature invariant subspace. A binary image of a traffic sign is used as a feature extraction object, the principal component analysis method and the linear discriminant analysis method are combined, and firstly the principal component analysis method is used for conducting feature extraction on the image of the traffic sign so as to obtain a feature matrix with the optimal description effect; then, the linear discriminant analysis method is used for conducting secondary feature extraction on the matrix so as to obtain a feature matrix with the optimal classification effect, and therefore features extracted in the traffic sign recognition method have the optimal description performance and the optimal discrimination performance; finally, the minimum distance classification method is adopted for identifying the extracted features, and tests verify that the traffic sign can be recognized accurately.

The invention discloses an overall kinetic parameter identification method based on a series robot, and belongs to the technical field of robot control. The overall kinetic parameter identification method comprises the steps: DH parameters of all axes of the robot are obtained, and a robot kinetic model is established; a minimum inertia parameter set of the robot is determined, and the robot kinetic model is linearized and converted into a form of a product of a minimum inertia parameter vector and a linearized matrix; an excitation track used for identification is calculated, data are collected and filtered, and a corresponding collecting moment and measurement matrix are obtained; and the collecting moment and the measurement matrix are normalized, weighted calculation is conducted on anoptimization algorithm, and kinetic parameters are obtained. According to the overall kinetic parameter identification method based on the series robot, various problems, such as the problem, causedby inconsistency of series axis scales, of excessive error of identification precision of overall or certain axis parameters, the problem of collecting data imperfection and the problem of data credibility, in overall parameter identification can be solved; and meanwhile the kinetic parameters of the robot can be identified on line, and the reliability of parameter identification can be greatly improved.

The invention relates to a method for permanent-magnet synchronous motor (PMSM) parameter on-line identification. The method is based on a speed sensorless vector control system of a virtual dc-qc coordinate system. The method comprises: 1) for a non-salient PMSM, establishing a qc-axis steady-state equation considering motor parameter error influence, the equation being u<qc> is approximately equal to R<s>i<qc>+ [omega]L<d>i<dc>+[omega][phi], wherein R<s> is the resistor of a stator winding, L<d> is an direct axis inductor, the [phi] is a magnetic linkage, [omega] is rotor angular velocity, u<qc> is qc-axis voltage, i<qc> is qc-axis current, i<dc> is dc-axis current; 2) and properly adjusting motor operating states repeatedly, and recording the steady-state value of each variable in the steady-state value, substituting into the equation to obtain a plurality of groups of steady-state relations, and obtaining the identification values of motor parameters through calculation of state variable increment. The method just duly adds short-time small disturbance of i<dc> and [omega] instructions in an original control strategy, having not influence on original control flow and system normal operation. The method is simple and fast in algorithm, and can realize non-salient or salient-pole PMSM parameter on-line identification in low cost.

The invention discloses a lithium ion energy storage battery SOC online estimation method, which comprises the steps of obtaining rated parameters of a to-be-detected lithium ion energy storage battery, and establishing an equivalent circuit model of the to-be-detected lithium ion energy storage battery, identifying model parameters of the established equivalent circuit model on line, establishingan improved adaptive extended Kalman filter of the lithium ion energy storage battery according to the established equivalent circuit model and model parameters obtained by online identification, andinputting the model parameters obtained by online identification into an improved adaptive extended Kalman filter of the lithium ion energy storage battery, and carrying out online estimation on theSOC of the lithium ion energy storage battery. Parameters of the first-order RC equivalent circuit model are identified in real time through a recursive least square method with a forgetting factor, and then the parameters are input into the improved adaptive extended Kalman filter, so that accurate estimation of the SOC of the lithium ion energy storage battery is completed.

This invention discloses a method for identifying on-line SFP electric modules, which connects the first and second test pins of a SFP module in short-circuit, when a control unit of a monoboard the SFP module places tests that the module has been plugged in the monoboard, said method includes the following steps: A, the control unit transmits a preset test signal to the first test pin of the SFP module by the pin of its own transmitting test signals, B, the control unit receives the fed back signals from the second test pin of the module by its pin receiving test signals and judges if said fed back signal is the preset signal, if so, it judges that the SFP module plugged in the monoboard current is an electric module, otherwise, it's not.

The invention discloses a multi-parameter identification method of a flux linkage adjustable permanent magnet auxiliary synchronous reluctance motor, and the method mainly comprises four steps: A, building a static inductance data table and an iron loss current data table based on off-line finite element electromagnetic field analysis under the cooperation of a torque flux linkage control algorithm; B, performing the high-frequency equivalent resistance and dynamic inductance identification based on high-frequency components; C, performing stator resistance and permanent magnet flux linkage identification based on fundamental frequency components; D, performing flux linkage separation of different coercive forces and rotor temperature identification. According to the method, the online identification of multiple physical parameters such as static/dynamic inductance, the stator resistance, permanent magnet flux linkage and rotor temperature of the flux linkage adjustable permanent magnet auxiliary synchronous reluctance motor can be realized, and the dynamic separation of low-coercivity permanent magnet flux linkage can also be realized.