57results about How to "Improve self-learning ability" patented technology

The invention discloses an intelligent fault diagnosis system for an ICNI system, which can improve the maintenance efficiency, carry out intelligent and automatic diagnosis and is applicable to the ICNI system. According to the technical scheme of the invention, a knowledge base and a management module thereof carry out standardization research and mathematical modeling based on a fault tree, an SQL Server database software framework is adopted, a relational database is used for building a logic relation among a fault phenomenon, a fault mode, a detection method, a historical case and a fault tree internal event to form the knowledge base; and a diagnosis information acquisition module interacts with an automatic testing system via Ethernet to acquire diagnosis data from the ICNI system and a testing instrument, a reasoning machine module adopts CBR and RBR hybrid diagnostic reasoning, after comprehensive judgment is carried out on the fault phenomenon inputted by the user, the field knowledge stored by the knowledge base and the diagnosis data from the automatic testing system, a reasoning method is automatically selected to carry out reasoning diagnosis on the fault, a reasoning process and a reasoning result are outputted to an explanation machine module, and a diagnosis report is generated.

The invention discloses a robust controller of a permanent magnet synchronous motor based on a fuzzy-neural network generalized inverse and a construction method thereof. The construction method of the invention comprises the following steps of: combining an internal model controller and a fuzzy-neural network generalized inverse to form a compound controlled object; serially connecting two linear transfer functions and one integrator with the fuzzy-neural network with determined parameters and weight coefficients to form the fuzzy-neural network generalized inverse, serially connecting the fuzzy-neural network generalized inverse and the compound controlled object to form a generalized pseudo-linear system, linearizing a PMSM (permanent magnet synchronous motor), and decoupling and equalizing the linearized PMSM into a second-order speed pseudo-linear subsystem and a first-order current pseudo-linear subsystem; and respectively introducing an internal-model control method in the two pseudo-linear subsystems to construct the internal model controller. The robust controller of the invention has the advantages of overcoming the dependence and local convergence of the optimal gradient method on initial values and solving the problems of randomness and probability caused by using the simple genetic algorithm, obtaining the high performance control, anti-disturbance performance and adaptability of the motor and simplifying the control difficulty, along with simple structure and high system robustness.

The invention relates to an environment quality detection control system and method. The system comprises an environment acquisition unit, a voice recognition unit, a network access unit, a data analysis unit and an environment quality adjustment unit. The environment acquisition unit carries out environment data acquisition and generates environment data. The voice recognition unit recognizes human voice, generates voice data and carries out simulation voice playback. The network access unit accesses the Internet, and acquires meteorological data and air quality data on the Internet, and uploads the voice data and corresponding command and environment data to the Internet cloud for storage. The data analysis unit carries out comprehensive analysis on various data and generates control data and voice feedback data. The environment quality adjustment unit adjusts the quality of indoor environment. According to the invention, environment quality detection, environment quality control, network access and voice recognition can be carried out; a user interacts with a device through a language; and the user can easily understand the outdoor environment quality for trend prediction and environment data statistics.

The invention provides an online advertisement recommending system and method for large-scale medium data and relates to the technical field of the calculation advertisement science. An advertisement dispatch engine module in the online advertisement recommending system is respectively connected with a user side, an advertisement management module and a flow analysis module. Parameter exchange is carried out between the flow analysis module and an advertisement searching module, a user behavior inquiry module and a webpage management module. A user behavior mining module is respectively connected with the advertisement management module and the user behavior inquiry module. The advertisement management module is connected with the advertisement searching module. According to the online advertisement recommending method, when a user finish accessing a webpage, the user is identified according to user information, user interests are inquired, user behaviors are learned, matched advertisements are searched for according to the predicted user behaviors, and finally the online advertisements are recommended to the user. The system has the good self-learning ability, can effectively improve the intelligent level of advertisement recommendation, and is suitable for online advertisement recommendation under the background with the large-scale data.

The invention discloses an intelligent ship environment threat target perception system and method. The system comprises a sensor module, a target recognition module, a comprehensive control unit, a short-distance recognition system judgment module and a target feature database. The sensor module comprises a navigation radar, a GPS / Beidou positioning navigation device, a marine AIS receiver, a small target radar, a high-definition video camera tracking device, a three-dimensional laser radar, a millimeter wave radar and a marine pickup device. According to the invention, the active and passivesensors from far to near configured by the intelligent unmanned ships are reasonably configured, the radar signals, photoelectric signals and audio and video signals are considered simultaneously torealize the threat target perception, identification and tracking functions, and the threat objects can be quickly, accurately and reliably identified. The system and the method have a continuous andimprovable self-learning capability, and along with the collection and enrichment of a target audio and video feature library, the capability and efficiency of threat target perception and target recognition of the intelligent unmanned ships can be continuously improved.

The invention relates to an improved particle swarm algorithm and the application of the improved particle swarm algorithm. The improved particle swarm algorithm includes the following steps that firstly, the algorithm is initialized; secondly, the positions x and speeds v of particles are randomly initialized; thirdly, the number of iterations is initialized, wherein the number t of iterations is equal to 1; fourthly, the adaptive value of each particle in a current population is calculated, if is smaller than or equal to , then is equal to and is equal to , and if is smaller than or equal to , then is equal to and is equal to ; fifthly, if the adaptive value is smaller than the set minimum error epsilon or reaches the maximum number Maxiter of iterations, the algorithm is ended, and otherwise, the sixth step is executed; sixthly, the speeds and positions of the particles are calculated and updated; seventhly, the number t of iterations is made to be t+1, and the fourth step is executed. By means of the improved particle swarm algorithm, at the initial iteration stage, the population has strong self-learning ability and weak social learning ability, and therefore population diversity is kept; at the later iteration stage, the population has weak self-learning ability and strong social learning ability, and therefore the convergence speed of the population is improved.

The invention discloses a control method of a household air conditioner. The control method is characterized by comprising the following steps: after the air conditioner is electrified, a user selects the air conditioner to enter an intelligent operation mode or a common manual set mode; when the air conditioner enters the intelligent operation mode, the air conditioner automatically selects a working mode operated for the most times in the latest week; after the whole air conditioner is operated, the air conditioner times the present operation mode; after the timed time t reaches the set time, the air conditioner counts the present working state once as the judgment proof of the use times of the common working state at the next starting time; and the timer is reset after counting for reclocking. The method customizes the special air conditioner control mode for users to adapt to the use habits of all the users so as to further improve the intelligence and the humanization of the household air conditioner.

This invention discloses one network intruding testing method and to one safety test method and to the test method to judge whether the network data flow to be tested is intruding, wherein the method comprises the following steps: pre-processing; studying step, testing step data process structure, wherein, the studying and testing steps adopts BP network as module for data processing and for studying or testing steps according to different system status.

The invention belongs to the field of control of an electro-hydraulic servo system, and relates to a force / position cooperative control method of an electro-hydraulic servo system. In the implementing process of the method, a position output signal and a force output signal of a valve control cylinder of the electro-hydraulic servo system in a work process are analyzed, outer-loop control of force is additionally provided as feedforward compensation based on position control, a PID controller and an adaptive fuzzy neural network controller are designed to respectively and individually control a position control portion and a force control portion, and cooperative control of the position signal and the force signal of the electro-hydraulic servo system is finally realized. The object of the invention is to reduce the vibration and the impact in the work process of the electro-hydraulic servo system due to stress and improve the positioning precision and stability of the system. The method includes steps: the position control portion measures the position output signal of the valve control cylinder through a displacement sensor and feeds back the position output signal to a position signal input portion, the position output signal is compared with an input signal, and a position error signal is obtained; the force control portion measures the force output signal of the valve control cylinder through a force transducer and feeds back the force output signal to a force input portion, the force output signal is compared with a force input signal, and a corresponding force error signal is obtained; and finally the error signal of the position control portion and the error signal of the force control portion are added (namely the force error signal is regarded as feedforward compensation) as a position expected input error signal of the whole valve control cylinder, the valve control cylinder dynamically adjusts the position signal and the force signal of the valve control cylinder by employing incremental control, and cooperative control of the position and the force of the electro-hydraulic servo system is finally accomplished.

The invention discloses a novel least square support vector machine modeling method for a thermal error of a numerical control machine. The method comprises the following steps of: (1) selecting a kernel function and determining parameters; and (2) according to a principle of the least square support vector machine, establishing a machine thermal error model. A compensating system in the invention is simple in structure and reliable in application; and by means of the least square support vector machine modeling method, the model precision and the generalization capability are improved, and the defects of low precision, low generalization capability and the like of the conventional predicting method are overcome.

The invention provides a neural-network-based liver cancer auxiliary diagnosis system, method and device, and a medium. The neural-network-based liver cancer auxiliary diagnosis system, method and device convert explicit rules corresponding to the medical knowledge into implicit rules including connection weights by storing medical knowledge and a learning algorithm based on neural networks, and provide corresponding treatment plans for different diseases according to the implicit rules. The neural-network-based liver cancer auxiliary diagnosis system, method and device can improve the diagnosis of liver cancer simply and efficiently, can ensure the reliability of the diagnosis, and can effectively solve the limitations of doctors in the diagnosis and treatment of liver cancer due to doctors' inadequate and untimely grasp of standard diagnosis and treatment of liver cancer and interdiscipline knowledge, thus greatly improving self-learning ability of the system, and realizing constantself-improvement in the process of running.

The invention discloses a synthetic-aperture-radar (SAR) image change area detection method of self-paced learning. The method mainly solves the problem that in the prior art, sensitivity to speckle noises of synthetic-aperture-radar (SAR) images is prone to cause a loss of part of texture information of the synthetic-aperture-radar (SAR) images, and includes the following specific steps: (1) reading in the synthetic-aperture-radar (SAR) images; (2) carrying out normalization; (3) constructing a change detection matrix; (4) selecting a training sample set; (5) training a deep belief network; (6) constructing a probability matrix; (7) updating the probability matrix; and (8) obtaining a change detection image. According to the method, local information of the original images and self-learning ability of the deep belief network are effectively utilized, the speckle noises are reduced, the local information of the images is retained, and precision of change detection is improved.

The invention aims at providing a teaching method for learning English by combining with Chinese pinyin. The method uses the comparability between the Chinese pinyin and English; during the process for teaching English, the method uses Chinese phonetic symbol and consonant and vowel pithy formula to learn English to ensure that the student divides and marks syllables of English word and sentence, reads the pronunciation of the word and the sentence exactly and remembers the pronunciation in mind; after repeated learning, the student pronounces accurately, grasps the pronouncing rule for learning English by Chinese phonetic symbol and also spells and memorizes new words according to the rule, thereby improving the accuracy of pronouncing greatly and increasing vocabulary; the method changes the existing habit of most people who learns English, only can read, but can not pronounce and do not like to pronounce; and the method improves the sensitivity towards sound and the listening comprehension greatly, avoids deaf-and-dumb English, improves the ability of learning by self and increases the confidence and the interest of the student for learning English.

The invention discloses a synthetic aperture radar (SAR) image change region detection method based on neighborhood ratio and self-stepping learning, and mainly solves the problem that the prior art is sensitive to speckle noise of an SAR image, and the problem that part of texture information of the SAR image is lost easily. The method comprises the specific steps of (1) reading an SAR image; (2)normalizing; (3) calculating a neighborhood ratio difference value; (4) constructing a difference value matrix; (5) selecting a training sample set; (6) training a deep belief network; (7) constructing a probability matrix; (8) updating a probability matrix; (9) obtaining a change detection image. According to the method, the local information of an original image and the self-learning capabilityof a deep belief network are effectively utilized, and speckle noise is reduced, the partial image information is kept, and the precision of change detection is improved.

The invention discloses a primary and secondary school student word listening system based on symbol labeling, relates to the technical field of intelligent education, and aims to solve the problem that teachers and parents need to accompany in a traditional dictation education form to consume a large amount of time and energy. According to the technical scheme, new words are obtained through theimage processing technology, and a listening word database is generated; the speech recognition technology is used for obtaining an instruction, the word listening system plays vocabulary speech in the word listening database under the instruction, the monitoring module is used for reducing the possibility of cheating possibly occurring in the dictation process, and students can independently complete dictation tasks without the help of other people. According to the invention, the effect of greatly simplifying the burden of teachers and parents is achieved, and the self-learning ability of students is formed.

The invention discloses a telephone number identification method and device, computer equipment and a computer storage medium. The telephone number identification method comprises the following steps:acquiring telephone identifiers of all first-class calling numbers; extracting first-type calling numbers of which all telephone identifiers are junk telephones and forming a number set; extracting frequency spectrums from the fourth digit to the seventh digit of all telephone numbers in the number set; and judging the second type of calling numbers as junk calls according to the frequency spectrum. According to the embodiment of the invention, the number characteristics of the junk phone can be effectively and efficiently identified by constructing the frequency spectrum of the number of thejunk phone, so that the junk phone judgment speed and accuracy are improved.

The invention discloses a remote sensing image target detection model establishment method based on context enhancement and an application, and belongs to the technical field of image processing, andthe method comprises the steps: building a to-be-trained target detection model based on a neural network, and carrying out the target detection and training on a remote sensing image, obtaining a remote sensing image target detection model based on context enhancement; in the target detection model, using each module for extracting a multi-scale feature map F of the remote sensing image; extracting global context information of the F to obtain M; respectively enhancing boundary information and category information in the F to obtain M<E> and M<E><cl> and respectively capturing informationassociation between channels in M<E> and M<E><cl> to obtain channel weights W<d> and W<c>; and fusing the M and M<E> according to the W<d> to obtain a boundary information enhanced feature map F<E>, fusing the M and M<E><cl> according to the W<c> to obtain a category information enhanced feature map F<E><cl>, and fusingF, F<E>, and F<E><cl> to obtain a feature map F<E><ct>, and carrying outtarget detection on the feature map F<E><ct>. The method can improve the target detection precision of a remote sensing image.

The invention relates to a transformer substation fault diagnosis method and diagnosis device based on improved case-based reasoning, and the method gives consideration to the correlation between thecondition attributes of all rules in a case library and the correlation between the condition attributes of other rules, and combines with a similarity measurement method of a weighted Euclidean distance. The method can continuously update the case library, improves the adaptability and accuracy of reasoning through the correlation between signals, has the self-learning capability of the system, and effectively improves the work efficiency of operation and maintenance personnel.

The invention relates to a board thickness intelligent control method based on active learning, which belongs to the field of intelligent control technology. Self-learnable performance of a nerve network is used as a theoretical basis. A dynamic nerve network is combined with active learning; the parameter of a PID controller is adjusted in an online manner; and a development model based on active learning is constructed, thereby establishing an intelligent control system for thickness of band steel, so that the board thickness control system can perform self adjustment at proper time, and the control performance of the board thickness control system is improved through continuous training of the dynamic nerve network. The board thickness intelligent control method has functions of providing a mathematical model with high generalization capability and wide application range for online control parameter adjustment of the system; combining active learning with the dynamic nerve network, and improving self-learning capability of the network through active learning and acquiring network training samples, thereby improving adaptive capability of the system and realizing intelligent in a real meaning.

The invention discloses a virtual simulation teaching practical training system based on an AR technology, and the system comprises a learning module which is used for enabling a user to rapidly know the operation method of the teaching practical training system, and recording the learning condition of the user; and the AR data acquisition module is used for acquiring teaching tool parameters in an actual teaching scene and evaluation preference and character parameters of evaluation personnel based on the actual teaching scene and the teaching achievement evaluation scene scanned by the AR scanning terminal, and generating three-dimensional visual data from the teaching tool parameters. According to the invention, the learning module, the client sub-module, the introduction sub-module and the evaluation sub-module are arranged and cooperated together, and a recommendation algorithm is utilized, so that reminding is successfully realized according to the mastering condition of the user, and the operation process of mastering the system by the user is accelerated, so that the system is simpler and easier to learn; and the teaching and practical training efficiency of the user through the system is greatly improved.

The invention discloses a home cleaning robot control system based on adaptive strategy optimization. The home cleaning robot control system comprises a sensing system, a control system, a cleaning system, a driving system and an electric power system, wherein the sensing system collects environment information and electric power system information and transmits the environment information and electric power system information to the control system; the control system processes the collected information through an adaptive strategy control method according to the collected information, and transfers control signals to the driving system and the cleaning system. The home cleaning robot control system processes environment information obtained from a self-sensing system by adopting an adaptive strategy optimization method in the control system of a cleaning robot, then selects appropriate action, and transfers control signals of the control system to the driving system and the cleaning system through the sensing system, so as to ensure that the cleaning robot executes corresponding action to clean home environment.

The invention discloses a deep learning non-intrusive load monitoring method based on unsupervised optimization. The first part is to establish a supervised neural network deep learning model; the second part is optimization of the model by using an unsupervised learning mode, and the first part comprises the following steps: monitoring all load information in a period of time from a target load cluster; preprocessing the data by using an algorithm, and normalizing the data; performing neural network training on the preprocessed data; and evaluating a network training result. The second part is optimization of the model by unsupervised learning, iteration is carried out on each target load clustering center by utilizing a K-means clustering algorithm, a training data training model is reconstructed, a supervised learning algorithm is optimized by utilizing an unsupervised algorithm, and then analysis is carried out on power consumption behaviors. According to the non-intrusive load monitoring method provided by the invention, the self-learning capability, universality, sensitivity and accuracy of processing a non-intrusive load monitoring problem by using a deep learning algorithm are greatly improved.

The invention relates to an underwater vehicle docking control method based on reinforcement learning, belongs to the technical field of ocean control experiments, and introduces a reliable boundary updated by new and old strategies based on a PPO algorithm framework in deep reinforcement learning to improve the stability of agent learning. Meanwhile, a self-adaptive rollback cutting mechanism is adopted, the rollback strength is adjusted in a self-adaptive mode according to the situation that task successful completion experience is collected, and therefore the upper limit and the lower limit of new and old strategy updating are adjusted, and an intelligent agent is encouraged to explore in the initial training stage and converge stably in the later training stage. In the aspect of simulation training, a docking training environment considering ocean current and ocean wave interference is constructed, the training environment is used for intelligent body learning, and the anti-interference capability of the underwater vehicle is greatly improved.

The invention provides a high-power microwave weapon cooperative air defense system contribution rate evaluation method. The method comprises the following steps: constructing a fuzzy wavelet neural network; selecting a fuzzy wavelet neural network input index and preprocessing the collected index data; training a fuzzy wavelet neural network by using the preprocessed data; utilizing the trained fuzzy wavelet neural network to evaluate an incoming target, and respectively obtaining first target combat effectiveness and second target combat effectiveness; and determining the system contribution rate of the high-power microwave weapon in the cooperative air defense combat by adopting a system contribution rate solving method based on combat effectiveness increment according to the first target combat effectiveness and the second target combat effectiveness. According to the method provided by the invention, the comprehensive gain of the high-power microwave weapon for cooperative air defense combat is intuitively obtained.

The invention discloses a multi-modal medical image fusion method, which comprises the following steps of 1) source image decomposition: performing two-stage decomposition discrete stationary wavelet transform on at least two source images, 2) source image fusion: fusing the at least two source images through seven enhanced radial basis function neural networks, and 3) inverse wavelet transform: performing inverse wavelet transform, through the inverse wavelet transform, converting the seven fusion sub-bands into a fusion image. The method has the advantages that through translation of the invariant multi-scale transformation operator, enough direction information is provided, and the algorithm is relatively simple and easy to operate. In order to enhance the self-learning ability of the neural network, in combination with medical image fusion, pixel values, regional energy, pixel gradients and regional average gradients are adopted to form an input layer of the neural network, information of target feature points is accurately extracted, and meanwhile, information loss is avoided.

The invention discloses a simulation practice teaching system for Chinese medical meridian circulation, and is applicable to the field of Chinese medical teaching, in particular to a Chinese medical meridian circulation drawing practice method based on a modern information technology. The system solves the problems that existing Chinese medical meridian teaching processes are monotonous and abstract, students have troubles in understanding the abstract meridian concept, students' interest in learning Chinese medicine is difficult to arouse, and bidirectional teaching interaction between students and teachers cannot be achieved. The method mainly comprises the following steps of a, selecting a meridian name for a coming circulation practice, and entering a meridian circulation practice mode; b, utilizing a touch control interaction module to conduct drawing with fingers; c, by means of a storage module, conducting path comparison; d, comparing a user drawing path with a correct meridiancirculation path by means of a logic verification module; e, comparing a user path with the correct meridian path through a display module to give a judgement result of this practice. According to the simulation practice teaching system for the Chinese medical meridian circulation, abstract Chinese medical meridian theoretical acknowledge is converted into a direct and interactive simulation training system.

A nuclear reactor early warning system includes a detection unit which measures various parameters of a nuclear reactor through sensors and detectors, converts the parameters into electrical signals and outputs the electrical signals, an input unit which is connected to a detection device and stores the electrical signals of the detection device into a preprocessing unit, the preprocessing unit which filters data signals, discards some wrong data and then stores the rest data into a critical heat flux database, a calculation unit which calculates corresponding critical heat flux values in realtime according to the measurement data filtered by the preprocessing unit, a self-learning unit which calculates the data in the critical heat flux database and continuously corrects a calculation model according to a principle of minimum error, a feedback unit which returns a result of a self-learning process to the calculation unit, and an output unit which displays a critical heat flux resultobtained by the calculation unit.

The invention discloses a cloud edge collaborative document classification system and method based on deep reinforcement learning. The method comprises that a document keyword analysis module and a document abstract analysis module which obtain a document keyword and a document abstract according to a to-be-classified document; a machine document content classification module obtains a first classification tag according to the document abstract, the document keyword and the to-be-classified document; document classification personnel selects a document classification tag according to the document abstract, the document keyword and the first classification tag in a manual classification module to obtain a second classification tag; and a document classification efficiency evaluation modulecalculates a classification efficiency value according to statistical efficiency parameters, directly stores a classification result if the classification efficiency value is lower than a set threshold, and otherwise, takes an expert classification result as a final result. According to the method, the accuracy of text classification can be improved by combining manual classification and expert classification, the professional ability requirement of professional document classification on classification personnel is reduced, and the working efficiency of the classification personnel is improved.