704 results about "Online identification" patented technology

The invention provides a combined estimation method for lithium ion battery state of charge, state of health and state of function. The combined estimation method comprises the steps that the state of he---alth of a battery is estimated online: open circuit voltage and internal resistance are identified online by adopting a recursive least square method with a forgetting factor, the state of charge is indirectly acquired according to a pre-established OCV-SOC corresponding relation, and then the size of battery capacity is estimated according to cumulative charge and discharge electric charge between two SOC points; the state of charge of the battery is estimated online: the state of charge of the battery is estimated by adopting the Kalman filter algorithm based on a two-order RC equivalent circuit model, and the battery capacity parameter in the Kalman filter algorithm is updated according to the estimation result of battery capacity; and the state of function of the battery is estimated online: the maximum chargeable and dischargeable current is calculated based on the voltage limit and the current limit of the battery according to internal resistance obtained by online identification, and then the maximum chargeable and dischargeable function can be obtained through further calculation.

A computer method and apparatus for adaptive model predictive control (MPC) of multivariable processes is disclosed. The adaptive MPC system can perform automatic implementation for a new MPC controller, and, for an existing MPC controller, it can perform automatic maintenance when necessary. The adaptive MPC system consists of three modules: an MPC control module, an online identification module and a control monitor module. In MPC implementation, the online identification module and the MPC control module work together and perform various steps in MPC implementation automatically and efficiently. When the MPC controller is online, the control monitor module continuously monitors the MPC performance and model quality. When control performance becomes poor and considerable model degradation is detected, monitor module will start the maintenance by activating the online identification module. The identification module will re-identify the model and replace the old model. For strongly nonlinear process units, multiple models are identified and used in control.

The invention relates to an identification method for the pipeline safety based on BP network and used in a monitoring device for the fiber pipeline leakage, belonging to the pipeline monitoring technique. The method comprises the following steps: the voltage signal is acquired after the photoelectric conversion of two channels of interference light in the monitoring device for the interference typed distributed optical fiber pipeline by utilizing a synchronous data acquisition card with at least double channels; the eigenvector of the detection signal obtained by calculation is decomposed by the wavelet; a plurality of acquisition samples are selected for each intrusive event in order to train and test a BP neural network; after the model training is completed, the system can acquire the real-time testing signal, extract signal characteristics, import into the well trained model for the online identification of the types of the abnormal events occurring along the pipeline and locate the abnormal events. The invention has the advantages of simple and convenient method, small occupied resource of the system, good real-time quality, which can be realized by a plurality of forms.

The invention discloses an online simultaneous identifying method for load torque and inertia of a motor. According to the method, a model reference adaptive control inertia identification model is utilized to realize online identification of rotational inertia; then the rotational inertia acquired through identification is introduced into an extended Kalman filtering load torque identification model in real time as a control variable, so that decoupling identification of the inertia via the load torque can be realized. The method overcomes the defect of the present online parameter identification model for the motor that the online identification error is great due to intercoupling of the load torque and the rotational inertia, and can carry out accurate online identification to the load torque of the motor on the condition that the rotational inertia is unknown or the rotational inertia changes during operation of the motor. The method has the advantages of simple structure and minor calculation quantity, realizes decoupling of the rotational inertia via the online identification result of the load torque of the motor, and improves the practicability of the online parameter identification system.

The present invention discloses an SOC (state of charge) estimation method. According to the method, a battery OCV-SOC relationship module, a parameter acquisition module, an offline identification parameter value, a parameter discrete state space model, a battery parameter online identification module, a battery dynamic parameter update module and a battery SOC estimation module. The method includes the following specific steps that: 1, a discharge-standing experiment is performed, an OCV-SOC relationship expression is obtained through fitting, parameter values in an equivalent circuit model are identified; 2, a battery second-order RC system discrete state space model is established, and the battery model parameters are identified online and dynamically updated; and 3, the SOC of a battery is estimated online. With the method of the invention adopted, the defects of inaccuracy and accumulative error of the initial value of the SOC of a battery of in an ampere-hour integration method can be eliminated. The method is applicable to the dynamic change of the characteristics of the battery, can improve the accuracy of SOC online estimation and can be widely applied to the electric vehicles and storage battery management system field. The method has the advantages of high battery model precision, fast convergence, high stability and high reliability.

The invention relates to the technical field of PMSMs, solves a coupling problem during online identification of multiple parameters of a surface-mounted type PMSM, and achieves online decoupling identification of PMSM inductance, stator resistance and rotor flux linkage. Accordingly, the technical scheme adopted by the invention is that an online decoupling identification method of multiple parameters of the PMSM comprises the steps as follows: 1) identifying and coupling analysis of parameters of the PMSM; 2) a decoupling identification strategy, wherein voltage deviation before and after D shaft current injection is used for increasing the order of a motor mathematical equation, so that the decoupling identification of multiple parameters of the surface-mounted type PMSM inductance, the stator resistance and the rotor flux linkage are achieved; 3) neural network identifier design, wherein according to a parameter online identification problem of the PMSM, online identification is performed on motor parameters by adopting a self-adaptive neural network structure and a weight convergence algorithm based on a least mean square algorithm. The method is mainly applied to the design and manufacture of the PMSM.

Provided is a power battery SOC estimation method based on backward difference discrete model and a system thereof; the method comprises the following steps: step one, establishing a backward difference discrete model of a power battery, identifying parameters of the backward difference discrete model by a least square method including forgetting factors; step two, on the basis of the backward difference discrete model of the power battery obtained in step one, using self-adaptive extended Kalman filter in combination with a non-linear relationship between an open-circuit voltage and the SOC to complete an effective estimation of the power battery SOC. In the system, voltage and current sensors connected with the power battery are connected with an embedded microcontroller via an analog-digital conversion module. The microcontroller comprises a low-pass filter pre-processing module, a backward difference discrete battery model parameter online identification module, and an AEKF algorithm SOC estimation module. The obtained SOC result is sent to a CAN network of a display device. The power battery SOC estimation system based on backward difference discrete model is simple in structure; the parameter identification speed and precision are increased; the affection to the identification caused by history data is reduced; the calculation is convenient; and the SOC estimation precision is high.

The invention relates to a network application encrypted traffic recognition method and device based on protocol attributes, and belongs to the technical field of computer network service security. The device comprises an offline training module and an online identification module. The offline training module is composed of a data set obtaining module, a message type fingerprint establishment module based on a second order Markov chain and a certificate length clustering module. A training set is obtained through a data set obtaining module. Application fingerprints are obtained and stored according to the training set by the message type fingerprint establishment module based on the second order Markov chain; clustering results and application certificate cluster distribution probability are obtained and stored according to the training set by the certificate length clustering module. The offline training module is composed of a network traffic capturing module and a recognition module. The recognition module matches the network traffic obtained by the capturing module with a stored application fingerprint library one by one; moreover, the certificate clustering results are taken into consideration, thus obtaining a recognition probability; the recognition result is an application corresponding to the highest probability. Compared with the prior art, the method and the device have the advantage of improving the recognition accuracy and efficiency.

The invention discloses an instrument pointer image identification method of a high-robustness rod, mainly comprising two phases including offline calibration and online identification. The offline calibration phase comprises the following steps of: collecting images, pre-treating the images, marking and learning, calibrating an instrument position, learning instrument characteristics, selecting a pointer identification method and parameters, establishing a database and storing a calibrated result. The online identification phase comprises the following steps of: collecting the images, pre-treating the images, positioning an instrument, adjusting the gray level of the instrument images, extracting the pointer and calculating readings of the pointer. The instrument pointer image identification method of the high-robustness rod, provided by the invention, is high in robustness, and has stronger robustness to the translation, the rotation and the shielding of the instrument, instrument deformation caused by light changes and different viewing angles, the change of an imaging size caused by different distances, and the like; and the pointer readings can be more accurately identified and the stable work under an actual environment is realized.

An online identification document management service includes a host server having a web-based interface adapted to facilitate secure customer access to the host server. The host server is configured to receive identification data communicated from the customer through use of a computing device. The identification data includes at least one of an imaged identification document and information related to the imaged identification document. A processing module is in operative communication with the host server and is configured to process the identification data communicated from the customer. A processing software application is trained to classify the processed identification data and selectively extract data therefrom based on the classification. The processing software application is configured to selectively present the processed identification data for a customer verification via the web-based interface upon at least one of an unsuccessful classification and an unsuccessful extraction of data. A storage device is in operative communication with the processing module and is configured to store the extracted data as metadata upon at least one of the customer verification and the extraction of data. The processing software application is cumulatively trained to subsequently classify the processed identification data successfully upon the customer verification thereof. A content management software application operates on the host server and is configured to process a customer request through the web-based interface to selectively generate at least one reproduction of the processed identification data.

The invention provides a method for identifying a forgery seal based on a feature line randomly generated by matching feature points. The method comprises the following steps: a step of off-line stage and a step of online identification stage to acquire an image of a seal affixed document containing a seal stamp to be verified by an imaging device; a step of preprocessing the image of the seal stamp to be verified, and a step of extracting an effective stamp image of the seal to be verified; a step of extracting a feature point of the seal image to be verified; a step of extracting the feather point of the image of the seal stamp to be verified, and a step of constructing a database containing location information and descriptor information of each feature point; a step of matching the features point of the seal to be verified with that of a reference seal; a step of randomly generating image feature lines of an identifiable seal to be verified and the reference seal; and a step of evaluating truth or forgery of the seal to be verified. The stamps affixed with the same seal under different situations have the two characteristics of accordant number and distribution of the feature points and accordant image information. The method has the characteristics of simpleness, high efficiency, high identification accuracy and the like, and contributes to seal forgery prevention.

The invention discloses a power battery SOC estimation method based on a dynamic parameter model. The method comprises the following steps: performing a discharge-standing experiment to fit an SOC-OCV relational expression; carrying out a pulse discharge-standing experiment to identify a parameter value of a battery equivalent circuit model; establishing a battery system discrete state space model; performing online identification on battery parameters; obtaining the battery dynamic parameter model; and performing online estimation on the SOC of a battery. Further disclosed is a power battery SOC estimation system based on a dynamic parameter model. The system comprises a battery SOC-OCV relational expression determining module, a battery parameter offline identification module, a battery parameter online identification discrete state space model determining module, a battery parameter online identification module; a battery dynamic parameter determining module; a battery SOC estimation discrete state space model determining module and an SOC estimation module. The method and system provided by the invention improve the precision of a battery model, is lower in calculation complexity and can be widely applied to the field of an electric automobile power battery management system.

The invention provides an online identification method for a network multi-step attacking intention, wherein the method extracts characters of the received security incident warning message, carries out a pattern matching based on a character and attacking action happening sequence pattern collection. If the pattern matching is success, degree of association between fore and aft security incident warning messages according to the pattern matching is computed, and the network multi-step attacking relation between the security incident warning messages is recorded based on a compared result of the degree of association and a threshold value thereof. The invention filters the security incident with pattern matching but without association relation by computing the degree of association, and the real time online detect is realized by real time reporting a matching process and result.

The invention discloses a combustion optimization control method for a boiler. The combustion optimization control method is characterized by comprising the following steps: sampling a combustion nonlinear system of the boiler to obtain input / output data at the current moment; training the real-time sampled input / output data by an online incremental learning fuzzy neural network, building an online incremental learning predicting model of the combustion nonlinear system of the boiler; performing a nonlinear prediction control algorithm on the online incremental learning predicting model for realizing the optimization and the control of the combustion process of the boiler. According to the combustion optimization control method for the power station boiler of the online incremental learning fuzzy neural network, the nonlinear optimization problem in the predication control algorithm is solved by utilizing a particle swarm optimization algorithm through the online identification of the boiler combustion optimization model; the real-time optimization and control of the boiler combustion process are realized.

The invention discloses a time sequence and CNN-based unsafe behavior identification method and system. The method comprises the steps of inputting original data of a training set video of unsafe behaviors to a CNN for performing training and learning, and outputting a spatial eigenvector from the last pooling layer of the CNN; taking the spatial eigenvector as an input of a time recurrent neuralnetwork, learning a time sequence dependency relationship contained in the spatial eigenvector by using the time recurrent neural network to learn a time sequence representation of spatial behaviors,and obtaining an average pooling eigenvector; inputting the average pooling eigenvector to a softmax layer, thereby obtaining a deep mixed learning model of the CNN and the time recurrent neural network, namely, a softmax classifier; and performing online identification on a monitoring video of a construction site by utilizing the softmax classifier to obtain the unsafe behaviors of the construction site. The support can be provided for real-time investigation and correction of the unsafe behaviors in a whole building engineering construction process.

The invention protects a cross-updated active noise control system based on a novel algorithm for online identification of a secondary channel. The noise control system comprises 6 modules of noise signal filtering, momentum FxLMS algorithm, white noise generator, secondary channel modeling, main channel path and third adaptive filter update module. The object of the noise control system is to solve the problem that an active noise cancellation (ANC) system has a slow convergence speed and a small noise reduction amount in indoor noise-cancellation applications. The innovation lies in the factthat the unevenness of the power spectral density of a noise signal greatly affects the convergence speed of a control filter and a modeling filter in the indoor noise-cancellation applications. Themomentum FxLMS algorithm is proposed to update the weight of the control filter, and a variable step size LMS algorithm is used to update the weight of the modeling filter. The third adaptive filter using the Newton LMS algorithm is used to eliminate an error signal and the signal related to the reference input signal, and improves the modeling accuracy of the modeling filter and the convergence speed of the entire ANC system.

The invention relates to an emission monitoring time series data abnormal value detection method for a coal-fired unit. Specifically, all-working-condition historical sample data set of a measured point is subjected to working condition division by applying a k-means clustering algorithm according to unit operation states and environment-protection facility operating ways; then specific to each working condition historical sample data set, an abnormal value detection model is established based on a statistical hypothesis test method; and the time series data of the measured point is subjected to online identification by selecting different hypothesis test methods according to data distribution characteristics of the working condition sample set, and then abnormal points which are inconsistent with historical distribution and current change tendency can be positioned in time. Abnormal value online detection on the field monitored data is performed by the hypothesis test methods, and the abnormal points which are inconsistent with historical distribution or current change tendency can be dynamically identified, so that early warning for more complex fault monitoring is provided, refine degree of alarming is improved, and online monitoring management level and rapid response capability are improved.

The invention discloses a parameter identification method of a permanent magnet synchronous motor based on an inverter nonlinear compensation. Firstly, based on the vector control of the permanent magnet synchronous motor, the nonlinearity of an inverter is studied, the influence of inverter nonlinear factors on parameter identification is considered, and the online dead zone compensation is carried out for error voltage brought by nonlinearity. Then, the offline identification of parameters of motor closed-loop control is carried out based on dead zone compensation, and the identification accuracy is improved. Finally, the multi-parameter simultaneous online identification is achieved with the combination of a least squares method and a model reference adaptive method. A motor inductance,a permanent magnet flux linkage and a resistance value can be identified at the same time for a short time at a low-speed condition, and the method has great popularization and application values.

The invention discloses a method for intelligently controlling the pressure difference of a cement raw meal vertical mill. In the method, internal pressure difference of the mill is taken as a major controlled variable, external circulating current is taken as an auxiliary controlled variable, a feeding amount is taken as a major control variable, and a water spray amount is taken as an auxiliarycontrol variable. The method comprises the following steps of: judging a current production condition according to the internal pressure difference of the mill and external circulating current information with a production rule method; and under the normal working condition, controlling the internal pressure difference of the mill by adopting GPC-PID (Generalized Predictive Control-Proportion Integration Differentiation). In the method, a mathematic model for describing the change of the pressure difference along with the feeding amount can be established only by using input and output data of a process without analyzing the complex material milling process of the raw meal vertical mill, the identifying process is simple, and online identification is available; the method has certain adaptability to the change of the production condition; and the advantages of multistep prediction, rolling optimization and feedback correction of a GPC algorithm as well as the advantages of easiness and high robustness of a PID algorithm are fully utilized, so that the labor intensity of field operating personnel is lowered greatly.

The invention discloses a non-contact conductive gas / liquid two-phase flow pattern identifying device and method thereof. The device comprises a gas / liquid two-phase flow measuring pipeline, an alternating-current excitation source, a capacitive coupling non-contact conductive sensor, a temperature sensor, a current / voltage conversion unit, an alternating-current rectifying unit, a filter unit, a direct-current amplifying unit, a data acquisition module and a computer. The method comprises the following steps: firstly, the capacitive coupling non-contact conductive sensor obtains conductance signals of flows of the gas / liquid two phases; then, characteristic values of the obtained conductance signals are extracted via a time domain collection method and a Fourier transformation method; and finally, the characteristic values are input to a flow pattern classifier for online identification. The device and method provided by the invention can effectively prevent reduction of the conductivity measurement performance caused by polarization effects and electrochemical corrosion effects, and improve the reliability and robustness of a flow pattern identification system; furthermore, the device is simple in structure, easy to mount, low in cost and the like. The invention provides a novel effective path for measuring parameters of the gas / liquid two phases in the fields such as energy source, petrochemical industry and the like.

The invention discloses a target identification method based on three-dimensional point cloud data. The target identification method comprises the following steps of: carrying out external cuboid obtaining on three-dimensional point cloud data of a marked category; extracting a regular geometry from the three-dimensional point cloud data of the marked category; constructing a model database; constructing a model data table; and inputting point cloud data of a target to be identified and carrying out online identification. A target identification system based on the three-dimensional point cloud data comprises a first module, a second module, a third module, a fourth module and a fifth module, wherein the first module is used for executing external cuboid obtaining on the three-dimensional point cloud data of the marked category; the second module is used for executing extraction of the regular geometry from the three-dimensional point cloud data of the marked category; the third module is used for executing construction of the model database; the fourth module is used for executing construction of the model data table; and the fifth module is used for executing input of the point cloud data of the target to be identified and online identification. The method and the system are convenient and rapid in the target identifying process, and are high in accuracy. The method and the system are widely applied to the technical field of mode identification.

The invention relates to an online identification method on the rotor resistance of an asynchronous motor, and the method is used for optimizing resistance parameters in the vector control with a speed transducer. The invention includes the following procedures of collecting the electromagnetic torque Te and the psi 2parameter of the flux linkage of a rotor, identifying the rotor resistance according to the electromagnetic torque Te and the psi 2 parameter of the flux linkage of the rotor by the formula that r2 is equals to 3 / 2Pn omegas psi <2>2 / Te, among which the p n is the number of the pole-pair of the motor, and the omega s is the slip angular frequency, taking the identified rotor resistance as the parameter to implement the vector control of the motor. The online identification method of the invention can identify the rotor resistance of the asynchronous motor on line on the vector control of the motor with the speed transducer by a voltage model method, thereby being simple and practical, and having higher precision. The invention can identify the resistance parameter of the rotor on line accurately under any condition that load exists, and the resistance parameter is stable. Experiments prove that the invention can obtain right resistance parameter of the motor rotor accurately whether the motor is in cold state or in hot state.

The invention discloses a social media platform false information identification method. The method includes the steps that a false information sample set generation module uses a SGAN (Siamese Generative Adversarial Network) model to construct and generate a large-scale basic false information sample set and then uses a combination way of unsupervised learning and supervised learning to mark thefalse level of the generated basic false information sample; a false information offline identification learning module trains the identification of false information based on deep learning techniquesand iteratively optimizes parameters of a deep identification model; a false information online identification module firstly uses an exploration-to-use strategy to obtain candidate false informationset in newly released information and then conducts false identification and rating on the information in the candidate false information set in real time. Compared with the prior art, the method hasthe advantages of high accuracy, strong generalization ability, high speed and the like, and can be effectively applied to the fields of e-commerce, medical health, network information security and public opinion monitoring.

The invention discloses an apparatus for real-time online identification of strata lithology and identification method thereof, which relates to real-time online monitoring of lithology during the process of oil logging. The apparatus consists of a pulse generator (1), a pulsed laser (2), a sequence controller (3), a mesopore reflector (4), a holophote (5), a focusing mirror (6), a reaction chamber (7), a flow valve (8), a drilling liquid pool (9), a photoelectric receiver (10), optical fibers (11), a spectrometer (12), ICCD (13) and a computer (14). Cutting grains of drilling strata remained in drilling fluid are excitated to high-energy state by using focusing laser beam of high energy. During the transition of cutting grains from the excited state to the ground state, the components of cutting grains radiate their respective spectrums. By analyzing the characteristic spectrums, the elements and their corresponding concentrations of the samples are obtained and the lithology thereof are determined. The apparatus has the characteristics of simple structure, reliable method and easy operation, thus is suitable for real-time online detection of strata lithology during the process of oil logging.

An imaging system containing an electron-multiplying charge-coupled device detector and line-scan spectrograph is used for identifying wholesome and unwholesome freshly slaughtered chicken carcasses on high-speed commercial chicken processing lines. Multispectral imaging algorithms allow for real-time online identification of wholesome and unwholesome chicken carcasses.

The invention relates to a clustering analysis-based double-semi-dragging vehicle driving stability automatic identification and warning system. The clustering analysis-based double-semi-dragging vehicle driving stability automatic identification and warning system comprises a test data collecting module, a data pre-treatment module, an off-line clustering analysis module, an online identification module, a vehicle real-time data collecting module and a decision executing module. The output end of the test data collecting module is connected to the input end of the off-line clustering analysis module through a test data pre-treatment module; one output end of the off-line clustering analysis module is connected with the input end of the online identification module, and the other output end of the off-line clustering analysis module is connected with the decision executing module; the output end of the vehicle real-time data collecting module is connected with the input end of the online identification module through a real-time data pre-treatment module, and the output end of the online identification module is connected with the input end of the decision executing module. The clustering analysis-based double-semi-dragging vehicle driving stability automatic identification and warning system performs off-line clustering analysis on pre-treated collected test data, achieves classification of vehicle driving states through K-Means clustering analysis and then takes corresponding safety measures according to classification results.

The invention discloses a resistor-inductance load online identification algorithm applied to radio electric energy transmission primary side control. The system is divided into two stages which are a parameter online identification stage and a constant voltage power output stage. When a system works at a parameter online identification stage, the system needs to switch to work at two working frequency point, a first working frequency is set as an offset resonance frequency + / -5%, and a second working frequency is set as a resonance frequency f0. When the system works at the parameter identification stage, through feeding the instantaneous value of primary side current i1, the phase information and effective value or amplitude information of the primary side current are obtained and are used for calculating the reflection impedance of a main circuit, and thus the inductance and resistance values of the load and the coupling coefficient between primary and secondary side coils. After the corresponding inductance value and coupling coefficient are obtained, the system switches to the constant voltage power output stage, at that time, an output voltage control parameter is obtained according to the identified inductance value and the coupling coefficient, and the output voltage control parameter and the coupling coefficient are taken as initial values and are given to a controller.

The invention discloses an online monitoring analyzer and a monitoring system for sewage biotoxicity. The analyzer is in an integrated cabinet structure and the cabinet structure comprises an upper cabinet space and a lower cabinet space; the upper cabinet space is used for mounting a control system, a data collection and analysis system and a touch screen; the lower cabinet space is divided into front and back parts; the front and back parts are separated by a detachable separation plate; and the detachable separation plate is fixedly provided with a feeding pump, an air flow meter, a solenoid valve and a pipeline. The analyzer disclosed by the invention is in the integrated cabinet structure and is formed by an air-liquid conveying system, a wet type biochemical reaction system, an automatic control system and a data collection, processing and communication system; and the analyzer is used for online identification and early warning for the biotoxicity of the incoming sewage of a sewage treatment plant, prevents the sewage treatment efficiency from being influenced due to the reason that active sludge is poisoned by toxic components in the sewage, and has the characteristics of full-automation in operation, rapidness and accuracy in analyzing, simplicity in installation, convenience in maintenance and the like.

The invention relates to a multi-spectrum palm print ID certification method and special-purpose acquisition recorder; the multi-spectrum palm print image acquisition recorder comprises a digital video camera, a multi-spectrum light source, a light source controller, a communication interface, a computer, an instrument bracket with a base, an instrument housing and a palm placement plane, wherein, the digital video camera and the multi-spectrum light source are vertically mounted on the instrument bracket; the instrument housing is fixed outside the instrument bracket and the palm placement plane, so as to form a camera obscura; the light source controller is arranged on the interior side of the instrument housing; the digital video camera and the light source controller are connected with the computer through the communication interface; the invented instrument can acquire high-quality multi-spectrum palm print images, and has the advantages of real-time acquisition and free switch of light sources, so as to meet the requirements for online identification. The acquisition recorder and identification method provided by the invention can well improve universality and stability, as well as anti-fake property and identification performances of traditional palm print and vein identification systems.

The present invention discloses a conversation forecasting based online and offline identification switching method and system for robot. The method comprises the following steps that: the robot initiates a voice instruction; the voice instruction is matched with statements in a knowledge database of the robot; according to an answer mode of matching, engines are selected for identification; if an instruction with high identification rate needs to be obtained, offline identification is carried out; if an intelligent voice answer needs to be obtained, online identification is carried out; an identification result is output to the robot; the robot intelligently and selectively does the actions according to the identification result. Online identification and offline identification are combined for conversation forecasting; words or individual characters are identified by using the offline engine; and complete sentences are identified by using the online engine; so that the identification rate is increased; online and offline identification modes are effectively combined for achieving more effective communication between a user and the robot.