33results about How to "Realize self-learning function" patented technology

Provided are a computer-executed automatic error correction method for real-time court hearing speech recognition, a computer-readable storage medium, and a computing device. The method includes: discovering hot words and adding them to a hot word library in real-time in response to detecting a modification operation of a character recognized by a user in a courtroom system; carrying out real-timespeech recognition; and performing real-time hot word matching operation on the recognized text and performing hot word replacement operation. The court hearing speech recognition system according tothe embodiment of the present invention realizes the self-learning function of the court hearing system by automatically discovering hot words and constructing and updating a hot word library in realtime during the running process. Through ts_code technology in hot-word storage and hot-word substitution, different settings can be generated automatically for different accents through fuzzy soundconfiguration, so that the problem of speech recognition errors caused by multiple local accents can be easily and efficiently corrected.

The invention discloses a voice recognition method based on user behavior characteristics. Firstly user voice data are acquired by a voice recognition terminal and transmitted to a server to be processed and analyzed; the behavior habits and the sound characteristics of users are analyzed through combination of information of the users; the rules of grammar and the command word library further meeting the behavior habits of the user are generated according to different behavior habits of each user; a sound characteristic model is generated according to the sound characteristics of each user; and the rules of grammar, the command word library and the sound characteristic models of the voice recognition terminal are adjusted and updated so that the voice models are further optimized, the self-learning function is realized, each user is enabled to have more customized experience meeting his own behavior habits, the accuracy of voice recognition can be guaranteed and the reliability of voice recognition can be enhanced.

The embodiment of the invention discloses a PCA-BP-based emotion recognition method and system. The method comprises the steps that a first emotional sound spectrum feature vector of training audio isobtained by blending a first acoustic feature vector and a first sample entropy feature of the training audio; a PCA algorithm is adopted for conducting dimensionality reduction on the emotional sound spectrum feature vector, and a second emotional sound spectrum feature vector is obtained; the second emotional sound spectrum feature vectors of the training audio corresponding to various emotionsare input into a preset BP neural network model separately for training, and sound spectrum feature vector emotion models corresponding to various emotions are obtained; a second acoustic feature vector and a second sample entropy feature of to-be-detected audio are blended to obtain a third emotional sound spectrum feature vector of the to-be-detected audio, the third emotional sound spectrum feature vector is compared with all the acoustic feature vector emotion models, the emotion model matching degree is calculated, and the acoustic feature vector emotion model corresponding to the maximum emotion model matching degree is output. The technical problems that the current voice emotion recognition process is complex, the implement difficulty is high, the accuracy is low, and the efficiency is low are solved.

The invention relates to a method for adjusting welding parameters based on large step pitch calibration and local Newton interpolation, which comprises the steps of: automatically generating parameters of other working points in the interval by means of Newton interpolation with a plurality of calibration working point parameters with good welding effect to achieve continuous adjusting effect and supplement the deficiency of discretized segmented adjustment of a digital welding power supply; and raising welding parameter storage modes of different attribute priorities according to different parameter generating manners to optimize the welding parameters. By using the invention, parameters of various welding processes such as CO2 welding and P-GMAW (Pulse-Gas Metal Arc Welding) and the like are generated automatically and used for welding, so that the process is stable and the weld joint is formed well. The method can realize the coverage of the welding parameters, realize the automatic learning function of a digital welding machine so that the welding parameters are more favorable as being used further, thereby meeting the requirements of intelligentization of the digital welding machine.

The invention discloses a smart home interaction control method. The method comprises the following steps: a voice instruction of a user is acquired; the voice instruction is parsed to acquire equipment information and function instruction information; infrared code tables corresponding to the equipment information are loaded; and according to the infrared code tables, the function instruction information is compiled into infrared control instructions for controlling equipment corresponding to the equipment information. The invention further provides a device for implementing the method. The device comprises an acquisition module, a parsing module, a loading module and a compiling module. The technical scheme is achieved on the basis of infrared remote control and voice control technologies, is suitable for traditional household appliances and effectively improves the user experience.

The invention provides an Internet of Things sewing machine. The Internet of Things sewing machine comprises a sewing machine main body (1), a display screen (2), an electric cabinet (3), a Web data server (4) and a control terminal (5), wherein the display screen (2) is arranged on the sewing machine main body (1) and is used for identifying and displaying user information and sewing machine information; the electric cabinet (3) is connected with the sewing machine main body (1) and is used for controlling the operation of the sewing machine main body (1), receiving recorded information and transmitting the information to the Web data server (4); the Web data server (4) is used for receiving, processing and storing the information and transmitting the processed information to the controlterminal (5) and the electric cabinet (3); the control terminal (5) is used for identifying and managing person information, receiving and displaying the processed information and transmitting an instruction. According to the Internet of Things sewing machine disclosed by the invention, the sewing machine can autonomously upload sewing data; the control terminal synchronize in real time, so that the intelligent degree of the sewing machine is improved, the consumption of human resources is effectively reduced and the production efficiency is improved.

The invention discloses a traditional Chinese medicine assistance method based on the Internet. The method is achieved through a front end, a middle end and a rear end. The front end collects disease symptoms. The middle end conducts clinicopathologic analysis on the disease symptoms and determines the prescription. The rear end verifies and outputs the prescription.

The invention relates to an intelligent self-learning system for electromechanical equipment. The intelligent self-learning system is connected with the electromechanical equipment and used for auxiliary analysis of operation of the electromechanical equipment. The intelligent self-learning system comprises a data collecting unit, an operation model self-learning system and a standard model data base, wherein the data collecting unit is connected with the electromechanical equipment and collects multiple items of parameters in operation of the electromechanical equipment in real time; the operation model self-learning system is connected with the data collecting unit, builds a standard operation model according to the multiple stems of parameters in operation of the electromechanical equipment and refreshes the standard operation model; the standard model data base is connected with the operation model self-learning system and stores the standard operation model after each refreshing. The intelligent self-learning system can collect all items of information data in operation of the electromechanical equipment and obtain the standard operation model of the electromechanical equipment to achieve the self-learning function on the basis of the collected data, so that the intelligent self-learning system can provide reference for research personnel in the process of analyzing the operation situation and faults of the electromechanical equipment and is beneficial to performing parameter preset and analysis on the electromechanical equipment.

The invention relates to a digital globe, which comprises a globe body. Graphic elements of a variety of geographic information are distributed on the globe body; and different hidden identification codes are printed in various graphic element. The digital globe is also provided with a digital reading device for reading the hidden identification codes. The digital reading device comprises an image sensor, an image processing unit, a data storage unit, a speech decoding unit, a speech playing unit and a power supply unit for providing the working power supply. The image sensor scans the graphic elements of the globe body and transmits the corresponding hidden identification codes to the image processing unit; the image processing unit analyzes the hidden identification codes and transmits a corresponding signal to the data storage unit storing the globe information data; the data storage unit calls related speech data and transmits the related speed data the speech decoding unit for decoding; and the decoded speed data is played by the speech playing unit. The digital globe with the structure has rich information and provides convenience for self-study.

The invention discloses a program code verification method and device using a vulnerability model. The method includes: scanning a source code of a program; when a performance function is scanned, determining whether an input parameter of the performance function is transmitted or not to a dangerous function defined by the vulnerability model and whether a path where the input parameter is transmitted from the performance function to the dangerous function lacks or not a check function defined in the vulnerability model; if the input parameter of the performance function is transmitted to the dangerous function defined by the vulnerability model and the path where the input parameter is transmitted from the performance function to the dangerous function lacks the check function defined in the vulnerability model, adding the performance function as a new dangerous function into the vulnerability model, and determining a security vulnerability detected. The vulnerability model at least includes the dangerous function of the code scanned and definition information of the check function. Through the application of the method and device, code verification is more efficient and applicable.

The invention discloses a vehicle as well as a monitoring method and a monitoring system for tire pressure thereof. The monitoring method comprises the following steps: acquiring position informationof the vehicle; controlling a tire pressure monitor to monitor the tire pressure of the vehicle according to the position information, thus acquiring tire pressure information of the vehicle; judgingwhether to trigger an alarm function of the vehicle or not according to current standard tire pressure information and the tire pressure information of the vehicle, and saving the tire pressure information and corresponding position information. According to the monitoring method for the tire pressure of the vehicle, disclosed by the embodiment of the invention, whether to trigger the alarm function of the vehicle or not is determined through fixed point monitoring on the tire pressure of the vehicle and according to monitored tire pressure information, and the running reliability of the vehicle in a specified position can be favorably ensured; by saving the tire pressure information of a fixed point position, related analysis and processing can be favorably carried out on the tire pressure during running of the vehicle.

The invention relates to an intelligent continuous casting tundish induction heating control method with a self-learning function. The method comprises the steps of parameter setting, induction heating input, natural temperature drop calculating, input power calculating, power output inputting, weighted value calculating, a initial value, continuous casting communication data, actual molten steeltemperature. Due to model parameter setting, different steel grade casting temperatures can be selected, calculation is conducted according to the molten steel casting temperature, the steel grade andthe continuous casting communication data, the natural molten steel temperature drop and tundish induction heating input power can be obtained. By means of the method, the casting temperature of a tundish can be accurately controlled, constant-temperature casting is achieved, and prejudgment is conducted on molten steel temperature drop.

The invention discloses a self-learning method and device for a hydraulic torque converter, electronic equipment and a storage medium. The self-learning method for the hydraulic torque converter comprises the steps that a sliding friction point oil pressure adjusting value corresponding to the hydraulic torque converter is obtained; a sliding friction point oil pressure learning value corresponding to the hydraulic torque converter is obtained according to the sliding friction point oil pressure adjusting value; and an oil pressure command value for control of a gearbox on the hydraulic torqueconverter is obtained according to a sliding friction point oil pressure calibration value and the sliding friction point oil pressure learning value. Through the technical scheme, a self-learning function of the hydraulic torque converter can be triggered according to the state of the sliding friction point oil pressure adjusting value, the process of oil pressure adjustment according to the sliding friction point oil pressure adjusting value is omitted, and the hydraulic torque converter can achieve sliding friction control of an ideal state in time.

The invention discloses a bridge inspection vehicle frequency conversion control system. The bridge inspection vehicle frequency conversion control system comprises a frequency converter, a PID controller, an image acquisition module and a speed prejudgment module; the frequency converter is in communication connection with a motor and is used for conducting frequency conversion control on the motor; the PID controller is in communication connection with the frequency converter and a speed sensor and is used for conducting closed-loop control on output signals of the frequency converter according to detection signals of the speed sensor; the image acquisition module is used for acquiring image information in the advancing direction of a bridge inspection vehicle; and the speed prejudgment module is in communication connection with the frequency converter, the PID controller and the image acquisition module and is used for prejudging the driving speed of the bridge inspection vehicle according to the image information acquired by the image acquisition module on the one hand and used for correcting the driving speed prejudging process according to real-time control information sent by the PID controller. According to the bridge inspection vehicle frequency conversion control system, the defects in the prior art can be overcome, and the real-time computation amount of the PID controller is decreased by presetting the driving speed.

The invention relates to a method for making a virtual display course based on a material object teaching material, which comprises the following steps: determining key knowledge points needing classroom teaching or student mastery in a paper teaching material; preparing each key knowledge point into corresponding three-dimensional data; obtaining external information associated with the key knowledge points, and collecting the external information and the three-dimensional data into a three-dimensional data primary model; performing feature identification on each three-dimensional data primary model to prepare an AR code; and setting the AR code at a corresponding key knowledge point of the paper teaching material; setting an AR code or an AR total code at each key knowledge point of the paper teaching material; a user learning corresponding progress by scanning an AR code or an AR total code. The user can select different virtual scenes, real scenes or self-made information according to interests, the corresponding information and the three-dimensional data primary model or the actual data model are integrated into a three-dimensional data model, the three-dimensional data model can be fed back through the Internet, and the self-learning function of the three-dimensional data primary model is achieved according to feedback information.

The invention discloses a dynamically adjustable and controllable multimode interference coupler. The coupler comprises a coupler body. The coupler body comprises a supporting substrate, a coupler base body arranged on the supporting substrate, a plurality of input waveguides and a plurality of output waveguides, the input waveguides and the output waveguides are arranged on the coupler base body.And the coupler further comprises a dynamic regulation and control element used for regulating and controlling a refractive index of the coupler base body. Based on a MMI coupler, compared with an MZI array structure, the coupler has advantages that the coupler is simple in structure, and only one optical element (MMI) can achieve computing power which can be achieved by an array composed of dozens of MZIs; the coupler is insensitive to a temperature change and can be applied to occasions with the large temperature change; a bandwidth is wide, a design is more flexible, and applicability is stronger; and the coupler is insensitive to machining errors so that machining is simpler.

The invention provides an automatic reminding method and device and an electric automobile, and relates to the technical field of electric automobiles. The automatic reminding method is applied to a vehicle-mounted information entertainment system, and the vehicle-mounted information entertainment system is connected with a cloud server and an automobile body information collection device. The method comprises the steps of collecting external environment information of the electric automobile, driving parameter information of the automobile and user operation behaviors; according to the user operation behaviors, updating a automobile using habit currently stored in a user account; and generating reminding information according to the automobile using habit, the external environment information or the driving parameter information of the automobile. According to the scheme, the reminding information is autonomously provided for the user according to the collected information, and the automobile using experience and feeling of the user are improved.

The invention discloses a self-learning method and device of a hydraulic torque converter, electronic equipment and a storage medium. The self-learning method of the hydraulic torque converter includes obtaining the oil pressure adjustment value of the sliding friction point corresponding to the hydraulic torque converter; Obtain the corresponding sliding point oil pressure learning value of the hydraulic torque converter according to the sliding point oil pressure adjustment value; obtain the gearbox control hydraulic torque converter according to the sliding point oil pressure calibration value and the sliding point oil pressure learning value Oil pressure command value. Through the technical solution of the present invention, it is beneficial to realize the self-learning function of the hydraulic torque converter triggered according to the state of the oil pressure adjustment value of the sliding point, and the process of adjusting the oil pressure according to the oil pressure adjustment value of the sliding point is omitted, which is beneficial Make the hydraulic torque converter realize the slippery friction control in the ideal state in time.

The invention provides a charger. In order to solve the problem that the prior charger cannot carry out self-learning calibration on an intelligent battery, the charger is provided, wherein the charger comprises an adapter interface, a battery interface, a management module which comprises a control unit used for reading first information of the intelligent battery and sending a control signal according to the first information, a charging circuit for charging the intelligent battery when the control signal is a charge signal, a discharging circuit for discharging the intelligent battery when the control signal is a discharge signal and a hardware key for controlling the intelligent battery to enter a self-learning state, wherein the discharge circuit is constant load used for providing constant current or voltage for the intelligent battery. After the charger is adopted, a user can solve the self-learning process of the intelligent battery when using the charger and does not need to connect the intelligent battery to a notebook for self-learning.

The invention discloses an intelligent soot blowing control method, system and storage medium based on multi-dimensional evaluation factors. By selecting a plurality of boilers reflecting the contamination degree of the heating surface corresponding to the soot blower and the influence of soot blowing on the safety and economy of boiler operation The operating parameters are used as the evaluation factors for the selection of sootblowers, and each evaluation factor is prioritized and grouped according to the impact of each evaluation factor on the boiler safety, the degree of economic operation, and the degree of correlation with the contamination of the heating surface; Determine whether the value of each evaluation factor is within the normal state range, count the number of evaluation factors within the priority level of each group whose value is not within the normal state range, and calculate the number of evaluation factors within the priority level of each group whose value is not within the normal state range. The number of evaluation factors judges whether soot blower is needed for soot blowing, thereby reducing the soot blowing steam loss of the power plant, reducing the risk of blowing damage to the heating surface, and improving the safety and economy of boiler operation.

The invention discloses a smart home control method with self-learning and dynamic adjustment functions, which includes: a controller communicates with a wireless router to obtain access time period information of an access device in a set time period, and according to the access The time period information classification equipment is: mobile terminal, fixed terminal; the controller records the control parameters for controlling the home equipment within the set time period; the controller obtains the access time period information of the mobile terminal to generate home time period information; the controller Obtain control parameters according to the access time period information of the mobile terminal, and generate at-home mode information; the controller controls the home equipment to work to match the at-home mode information according to the at-home time period information; the controller acquires all mobile terminals offline for a set time value , the controller controls the work of the home equipment to match the information of the away-from-home mode. The invention provides a smart home control method capable of providing users with ultimate comfort and good energy-saving effect.

The invention discloses a personalized language offline learning method based on a deep neural network. The method comprises the following steps: S11, making a system enter a learning state, S12, by the language to be learned by the system, realizing user reading and by the system, continuously performing voice acquisition, performing deep neural network operation on the acquired voice data to obtain neural network acoustic features, and storing the features, S13, repeating the step S12 for multiple times, and performing deep neural network operation on the voice data acquired each time by thesystem to obtain neural network acoustic characteristics, calculating marginal distances of the plurality of obtained neural network acoustic features in pairs, and S14, by the system, judging the obtained marginal distance. According to the invention, the neural network acoustic features are used to realize personalized language self-learning recognition, each user can learn personalized language in an offline state of the device, and the function of customizing voice command words by the user personally can be realized.

The invention discloses an electric passenger door intelligent control system and an anti-pinching method. A center processor module, a door opening and closing input signal detection module, a motordetection module, a door position sensor, a temperature sensor, a motor driving module, a nonvolatile memory module, a communication module and a fault indicating module are included, and the door opening and closing input signal detection module, the motor detection module, the door position sensor, the temperature sensor, the motor driving module, the nonvolatile memory module, the communicationmodule and the fault indicating module are all connected with the center processor module. By means of the electric passenger door intelligent control system and the anti-pinching method, the safetyrequirement that a passenger door prevents pinching of personnel can be better met, and the problem that the anti-pinching performance is limited by the environment temperature is solved; the convenience requirement of installing and debugging of the passenger door is better met, and the problem that door opening and closing are not smooth because of different door opening sizes or the manufacturing tolerance is solved; the anti-pinching performance and door opening and closing smoothness reliability requirement in the design life period of a passenger door product is better met, and the problem that because of mechanical service life and abrasion, the anti-clamping performance and the opening and closing smoothness degrade is solved; and meanwhile, maintaining is easy, and the faults of an executing mechanism or an anti-pinching mechanism can be diagnosed and recorded.