Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
Hiro

33results about How to "Ability to learn independently" patented technology

Feedback fuzzy inference-based unmanned plane real-time route planning method

The invention relates to a feedback fuzzy inference-based unmanned plane real-time route planning method. The method comprises the following steps of: 1, updating a mode of the unmanned plane route; 2, establishing a fuzzy rule base; 3, acquiring a fuzzy antecedent; 4, designing a feedback fuzzy inference method; 5, defuzzifying; 6, determining other parameters; 7, determining the next position of the unmanned plane; and 8, completing the design. Compared with the conventional unmanned plane real-time route planning method, the method ensures that the unmanned plane has primary independent studying capability by applying the feedback fuzzy inference method so as to improve the route planning effect. The real-time route planning method designed according to the method has the advantages of short route, smooth path, global astringency and easy implementation, and is a more practical and reliable unmanned plane real-time route planning method for engineering application.
Owner:BEIHANG UNIV

Intelligent vein authentication method and system having autonomous learning capability

The invention discloses an intelligent vein authentication method and system having autonomous learning capability. The method comprises steps that model training of the full convolutional neural network depth learning model in repeated and regular global training and irregular local training modes is carried out, before online operation of the full convolutional neural network depth learning model, one global training is carried out through utilizing accumulated samples, and optimized characteristic extraction, characteristic identification and characteristic classification parameters are stored; after online operation of the model, newly-registered samples are utilized to carry out irregular local training of the model, and characteristic classification parameter optimization is completed; under the specific condition, the newly-accumulated samples are utilized to carry out regular global training of the model, and the characteristic extraction, characteristic identification and characteristic classification parameters are optimized for another time, the model is guaranteed to be the intelligent authentication system having the autonomous learning capability, and vein authentication precision is further enhanced.
Owner:通华科技(大连)有限公司

Indoor intelligent energy saving system based on Internet of Things and machine learning

The invention discloses an indoor intelligent energy saving system based on Internet of Things and machine learning. The indoor intelligent energy saving system based on the Internet of Things disclosed by the invention comprises a control input parameter module, a main control unit of the Internet of things, a terminal node set of the Internet of things, a terminal device set, an intelligent energy saving system module and an effect feedback module. The input parameter module comprises a variety of control input parameters, comprising, but not limited to, device control events, human sensingsensors, ambient light sensing sensors, area information, planned schedules, user preferences, scene information, environmental sensors, and the like. The indoor intelligent energy saving system disclosed by the invention has the beneficial effects that: 1. energy conservation, user experience and work performance can be comprehensively optimized by the cooperative use of the Internet of Things and an intelligent algorithm; and 2. multiple input parameters are comprehensively input, so that the decision making is more comprehensive and more scientific.
Owner:DATA TELLIT SUZHOU INC

Iris image quality classification method based on BP (back propagation) network and wavelet transformation

The invention discloses an iris image quality classification method based on a BP (back propagation) network and wavelet transformation. The method includes building a BP neural network model for iris image quality classification; selecting multiple iris image samples and subjecting the iris image samples to wavelet transformation to extract characteristic values of wavelet transformation coefficients; inputting the extracted characteristic values of the wavelet transformation coefficients into the BP neural network model for training; classifying iris image quality by the trained BP neural network model; inputting the characteristic values of the wavelet transformation coefficients, extracted from the iris image samples, into the trained BP neural network model; if an output result of the BP neural network model is as same as an expectation value of an available iris, judging the currently-input iris quantity to be available, or otherwise, judging the currently-input iris quantity to be non-available.
Owner:SHANDONG NORMAL UNIV

Unmanned driving system and control method thereof

The invention provides an unmanned control method, comprising the following steps of: S1, entering the unmanned mode, collecting real-time vehicle environment information with T for a cycle; S2, matching the collected real-time vehicle environment information with the existing vehicle environment information in the local database, and executing the corresponding automatic driving mode when the matching is successful, and proceeding to step S3 when the matching fails; S3, exiting the unmanned mode, recording the manual driving mode until the collected real-time vehicle environment information is successfully matched with the existing vehicle environment information in the local database; and S4, storing the failed real-time vehicle environment information and the manual driving mode in thenext T cycle in the local database as a new driving strategy. According to the unmanned driving system and control method thereof, the unmanned system is an artificial intelligence unmanned system that enriches the driving strategy in the database and has the ability to learn independently.
Owner:江苏萝卜交通科技有限公司

Method for conducting central air conditioner energy-saving regulation and control with artificial intelligence expert system

The invention discloses a method for conducting central air conditioner energy-saving regulation and control with an AI expert system. Through use of an AI learning module and a computer cloud servermodule, automatic adjustment is conducted, the autonomous learning capacity is achieved, and autonomous adjustment is conducted on a main area central air conditioner module. The method for conductingcentral air conditioner energy-saving regulation and control with the AI expert system is characterized in that the main area central air conditioner module, the computer cloud server module, an AI learning module and a time module are involved; when a time signal and a time reference signal are identical, the computer cloud server module transmits collected electric energy parameters to the AI learning module, and the computer cloud server module resets the time module and enables the time module to conduct timing again at the same time; when the main area central air conditioner module cannormally work according to a regulation signal, a feedback signal is output; and after the computer cloud server module receives the feedback signal of the main area central air conditioner module, the AI learning module stores the regulation signal. The method for conducting central air conditioner energy-saving regulation and control with the AI expert system improves the safety and reliabilityof a system and can prevent and completely eradicate generation and spreading of faults influencing normal operation of the system.
Owner:数方节能科技(烟台)有限公司

Single-emission multi-frame infrared target detection method

The invention relates to a single-transmission multi-frame infrared target detection method, which comprises the following steps of: starting from a feature pyramid network, describing the inequalityof contribution of each feature layer to fusion output on the basis of learnable weights, achieving bidirectional multi-scale feature weighted fusion between a low-resolution and strong-semantic feature layer and a high-resolution and weak-semantic feature layer, and constructing an auxiliary network; based on the cross-parallel ratio, considering the influence of an overlapping region and a non-overlapping region on a target function at the same time, constructing detector positioning loss keeping invariance to target scale change, constructing a target detector, and improving the sensitivityof a detection model to small target positioning errors. A VGG16 convolutional neural network is used as a feature extraction network and is integrated with an auxiliary network and a target detectorto form a single-shot multi-frame infrared target detection model fusing multi-scale feature weighted fusion and scale invariance positioning loss. The method has autonomous learning ability and highdetection rate, and is an effective way for solving the infrared imaging guidance target detection problem in a complex environment.
Owner:HENAN UNIV OF SCI & TECH

English learning management robot

The invention discloses an English learning management robot, which is a humanoid educational robot. The robot can interact with learners during teaching, and has a complete mechanism to stimulate the learners. If the learners do well in learning, the learners will be encouraged or prizes will be offered to the learners, thus increasing the learning interest of the learners. Through adoption of a regular learning unit, a regular learning function is provided to enable the learners to learn specified contents at the specified time, thus helping the learners to develop good learning habits. The robot communicates with the learners through voice recognition, voice processing and the like, and pure English communication better helps to improve the ability of the learners to speak English. An automatic learning ability is provided, the learners put books in a word scan identifying place, and the robot learns and understands contents of the books, so the robot can teach wider contents and is suitable for a wider range of population.
Owner:合肥慧动智能科技有限公司

Autonomous obstacle avoidance method based on deep learning and stereoscopic vision

InactiveCN113031597AImprove speed and precisionAvoid 3D reconstruction and path planning problemsImage enhancementImage analysisReal time acquisitionObstacle avoidance
The invention discloses an autonomous obstacle avoidance method based on deep learning and a binocular stereoscopic vision technology. Environment video information is collected in real time through a binocular camera, an obstacle three-dimensional detection algorithm is constructed based on the target detection in deep learning in combination with the binocular stereoscopic vision technology, and a robot performs autonomous obstacle avoidance according to an obtained obstacle avoidance steering decision, the three-dimensional coordinates of an obstacle and the obstacle avoidance angle. According to the method, deep learning target detection serves as a core, depth information of traditional machine vision is fused, the defect that only two-dimensional coordinate information can be obtained through a single deep learning algorithm is overcome, and complete three-dimensional coordinates are obtained. According to the invention, end-to-end learning in an obstacle avoidance scene is realized, complicated three-dimensional reconstruction and path planning problems are avoided, and the system cost is saved.
Owner:NANJING UNIV OF SCI & TECH

Visible light vein imaging method based on Monte Carlo

The invention discloses a visible light vein imaging method based on Monte Carlo, which belongs to the technical field of information perception and identification, and comprises the following steps:1) establishing a skin model, and carrying out optical model analysis on the skin; 2) carrying out skin imaging simulation on the basis of a Monte Carlo method; 3) carrying out skin tissue imaging, and finally realizing vein imaging. According to the method, a Monte Carlo method is used for simulating the color value of the skin image, and then the color value is used as the input and the skin parameter value is used as the output to train the neural network. For a given skin image, vein imaging can be realized from the obtained skin parameter value distribution map, the problems in the priorart are solved, vein imaging is performed for a visible light image of digital imaging, and an important foundation is laid for subsequent identification.
Owner:NANJING UNIV OF AERONAUTICS & ASTRONAUTICS

AI visual detection method and system based on deep learning, and readable storage medium

The invention provides an AI visual inspection method and system based on deep learning, and the method comprises the steps: obtaining an appearance surface image of a housing structure through an appearance surface visual inspection device, a first side visual inspection device and a second side visual inspection device of AI visual inspection equipment, obtaining a structural surface image of the shell structure through a structural surface visual detection device of the AI visual equipment; determining an appearance defect detection result of the shell structure according to the appearance surface image and an appearance defect detection model, and determining a structure defect detection result of the shell structure according to the structure surface image and a structure defect detection model; determining a quality detection result of the shell structure according to the appearance defect detection result and the structure defect detection result; wherein the appearance surface image at least comprises appearance surface images of the top surface of the shell structure and a plurality of side surfaces of the shell structure. According to the AI visual inspection method provided by the invention, the omnibearing inspection of the product is realized, and the outgoing quality of the product is improved.
Owner:中信云网有限公司 +1

Safety protection system and method for power edge computing

The invention discloses a safety protection system and method for power edge computing, and the system comprises a physical infrastructure safety protection module, a data plane gateway safety protection module, and a management safety protection module. By interface equipment and a micro-processor of the physical infrastructure safety protection module, identity registration and authentication identification are carried out on access equipment, automatic binding of interfaces of electrical equipment is carried out, real-time monitoring and sensitive data protection are carried out on attack behaviors through a data processor of the data plane gateway security protection module, and security protection of a data link is achieved. Password verification and log recording are carried out on platform operation and the like through a core processor of the management security protection module, so that the purpose of data security protection in the power edge computing system is achieved, and the data security and reliability of the whole power edge computing system are improved.
Owner:CHINA SOUTH POWER GRID ELECTRIC POWER RES INST

Intelligent atmospheric and vacuum system

The invention belongs to the field of petroleum refining, discloses an intelligent atmospheric and vacuum distillation system, and realizes digitization, automation, intelligentization and closed-loop operation optimization of an atmospheric and vacuum distillation device from planning to production operation. The system mainly comprises an atmospheric and vacuum AI model, a planning module, a communication optimization module and an intelligent execution module. According to the method, technologies such as artificial intelligence, big data, mathematical modeling and optimization, process integration and the like are comprehensively applied, and the atmospheric and vacuum AI model capable of accurately describing the atmospheric and vacuum dynamic production process is established on the basis of historical data of the production process. A planning module is established on the basis of the model, and a feasible material distribution decision is provided; meanwhile, a through disk optimization module is established on the basis of the model, and an optimal operation scheme decision is provided; and the intelligent execution module automatically executes the decision to realize closed-loop operation of the device. Intelligent integrated operation of the system is realized from planning, operation optimization and intelligent execution.
Owner:PROCESS INTEGRATION +1

Intelligent gateway for air conditioning operation and maintenance and maintenance detection method

The invention provides an intelligent gateway for air conditioning operation and maintenance, which is connected with external equipment. The intelligent gateway comprises an air conditioner and heating operation information interface module, an artificial intelligent analysis and processing module, a fault location collection and identification module, a wireless communication module, an abnormalalarm display module, a maintenance end module, a camera collection module and a data storage module. The invention further provides an air conditioning operation and maintenance detection method, which has the capability of independent study, the adjustment and repair of a central air conditioner are assisted, the failure of the air conditioner is prewarned in time, and the failure of the air conditioner is solved once by maintenance personnel with the help of the air conditioning operation and maintenance detection method.
Owner:ZHEJIANG UNIV

Intelligent reader implementation method and reader

The invention discloses an intelligent reader implementation method. The implementation method comprises the following steps: constructing a physical device for fixing a reading material and / or carrying out lamplight illumination on pages of the reading material; establishing a visual algorithm model to obtain a page content image of the reading material for recognition, and converting the page content image into voice information for output; establishing a dialogue understanding model to obtain user voice information for knowledge base matching, and outputting a matching result; acquiring anelectric signal and / or user voice information, and controlling the adjustment of the lighting, the page content image recognition of the book or the artificial intelligence dialogue. The intelligent reader provided by the invention has the functions of book fixing, visual reading, artificial intelligence dialogue and intelligent light adjustment; visual reading based on a deep learning network model can replace parents to read for children with limited reading ability, so that the children are limited by insufficient literacy, and the children have the ability of independently learning knowledge; and the system is multifunctional, high in intelligent degree and suitable for popularization and application.
Owner:小蒙(北京)智能科技有限公司

Feedback fuzzy inference-based unmanned plane real-time route planning method

The invention relates to a feedback fuzzy inference-based unmanned plane real-time route planning method. The method comprises the following steps of: 1, updating a mode of the unmanned plane route; 2, establishing a fuzzy rule base; 3, acquiring a fuzzy antecedent; 4, designing a feedback fuzzy inference method; 5, defuzzifying; 6, determining other parameters; 7, determining the next position of the unmanned plane; and 8, completing the design. Compared with the conventional unmanned plane real-time route planning method, the method ensures that the unmanned plane has primary independent studying capability by applying the feedback fuzzy inference method so as to improve the route planning effect. The real-time route planning method designed according to the method has the advantages ofshort route, smooth path, global astringency and easy implementation, and is a more practical and reliable unmanned plane real-time route planning method for engineering application.
Owner:BEIHANG UNIV

Diagnosis method of steam generator water supply system based on reinforcement learning of multi-source information

The invention provides a method for fault diagnosis of a steam generator water supply system based on multi-source information reinforcement learning, which includes monitoring the status of the steam generator, feed water pump and feed water valve in the steam generator water supply system, and obtaining multiple status information; The type clusters the associated state information to form a characteristic information cluster; establishes a characteristic matrix based on the integration of the characteristic information cluster, and extracts the state characteristic vector from the characteristic matrix; inputs the state characteristic vector into the data analysis model system to diagnose the fault type, and diagnoses the corresponding component Evaluate and regulate the state parameters of the system. Aiming at the multi-source information fault diagnosis of the steam generator water supply system, the present invention adopts technologies such as signal processing, mutual information correlation, nonlinear dimensionality reduction, deep reinforcement learning, etc., and solves feature extraction, feature fusion, feature integration, and feature learning of multi-source information It provides a new technical solution for the self-learning and state identification and diagnosis of multi-source information of steam generator water supply system.
Owner:WUHAN SECOND SHIP DESIGN & RES INST

Navigation signal recognition system and method

The invention discloses a navigation signal recognition system and method. The navigation signal recognition system comprises a self-navigation signal recognition module provided with a programmable controller module, an other-ship navigation signal recognition module provided with a signal processing server and a navigation signal display module provided with an upper computer software system. The navigation signal recognition system utilizes a programmable logic controller and the signal processing server and is combined with the upper computer software system to perform navigation signal recognition and display. The recognition method comprises the steps that through the programmable controller, navigation signals in a ship are collected, and a telescope is utilized to be combined witha camera to obtain the situation of ships on the periphery of the ship; through analysis and processing of pictures, navigation signals of the external ships are judged and recognized, finally, the collected navigation signals of the ship and external ships are sent to an upper computer for display, so that ship crew can understand navigational statuses of the ship and surrounding other ships, andreference proposals of safe navigation of the ships are provided, thereby improving the navigation safety.
Owner:SHENYANG INST OF AUTOMATION GUANGZHOU CHINESE ACAD OF SCI +1

Outbound robot distribution method

PendingCN113592250ASmall application limitationsAvoid relying too much on experienceForecastingCharacter and pattern recognitionPrincipal component analysisRegression analysis
The invention relates to an outbound robot distribution method, which comprises the following steps that: 1) when a customer calls, the customer selects to enter different incoming call skill groups through a routing module; 2) after entering the call-in skill group, dependent variables are predicted through principal component analysis and multiple regression analysis methods in sequence, i.e., the time required by the robot to dial the task load; and 3) based on the prediction result in the step 2), autonomous machine learning is carried out on the prediction result by using a method of combining a time sequence model and BP neural network weighting, an optimized learning rule is obtained, and then robot seats are distributed. Compared with the prior art, the method has the advantages that the labor cost is reduced, the prediction precision is improved, and the neural network autonomous learning ability is achieved.
Owner:交通银行股份有限公司太平洋信用卡中心

A Prediction Method of Rolling Mill Torsional Vibration

The invention relates to a prediction method of torsional vibration of a rolling mill. The prediction method comprises building a torsional vibration experiment platform and finding out an optimal prediction model of the torsional vibration of the rolling mill through real-time adoption of optimal kernel functions and parameters. The torsional vibration experiment platform is mainly formed by an alternating current frequency conversion motor, a gear reducer, a working roller which is arranged at the load end, a torsional vibration sensor, an elastic connecting shaft, a speed measuring sensor and an upper computer; the speed measuring sensor is used for measuring the speed which is transmitted to the elastic connecting shaft after the alternating current frequency conversion motor passes through the gear reducer, wherein the speed is used for representing the rolling speed of the rolling mill; the torsional vibration sensor is used for measuring the size of torque in the elastic connecting shaft; detected signals are transmitted by the two sensors to the upper computer through a serial port which is formed in the upper computer. According to the prediction method of the torsional vibration of the rolling mill, the torsional vibration prediction process is simple, the prediction accuracy is high, the objective evaluation standard of the torsional vibration is increased, the steps are clear, and the time and the labor are saved.
Owner:BEIHUA UNIV

Energy data analysis system and method based on big data

The technical field of data analysis of the present invention, in particular, relates to an energy data analysis system and method based on big data. The system includes: a data acquisition unit, including: various types of acquisition units deployed on site, and the acquisition unit is configured to collect raw energy data; a data storage unit: divides and classifies the collected raw energy data Storage, remove data redundancy; data preprocessing unit: check the integrity of data, remove noise and singular values, use data compression and transformation methods to reduce the number of effective variables or data samples to be considered according to the goal of the task, or Find the invariant identity of the data, and finally standardize the data. Through data segmentation and classification of energy data, it effectively reduces the redundancy of energy data and improves the efficiency of subsequent data processing; at the same time, it uses a deep learning model to conduct multi-scale analysis of energy data, improving the accuracy of data analysis results sex.
Owner:山东翰林科技有限公司 +2

Power distribution switch welding current diagnosis system and using method thereof

The invention discloses a power distribution switch welding current diagnosis system. The power distribution switch welding current diagnosis system comprises a comprehensive core management module, a data real-time acquisition module, a real-time curve drawing driving module, a real-time welding current analysis module and a current grading strategy management module; the comprehensive core management module is connected with the data real-time acquisition module, the real-time curve drawing driving module, the real-time welding current analysis module and the current grading strategy management module; the current grading strategy management module is connected with an instantaneous current waveform curve automatic screenshot module; and the comprehensive core management module is used for controlling the data real-time acquisition module, the real-time curve drawing driving module, the real-time welding current analysis module and the current grading strategy management module. According to the design, the intelligent degree is improved, and the automation degree is also improved.
Owner:DONGFENG AUTOMOBILE COMPANY

Similar application recommendation method, computing device and storage medium

The invention discloses a similar application recommending method, a computing device and a storage medium. The method includes: determining a first number of similar applications associated with a user identification according to a user identification of a current user and each application vector in a storage unit; The application features, industry type features, running sampling features and exposure times features of a number of similar applications are input to the ranking layer in the recommendation model for processing to output the second number of application identifiers; through the second number of application identifiers, obtain the first Application information of similar applications associated with the two number of application identifiers; the application information of similar applications associated with the second number of application identifiers is formed into an application list, and recommended to the current user.
Owner:BEIJING PARATERA TECH

Steam generator water supply system diagnosis method based on multi-source information reinforcement learning

The invention provides a steam generator water supply system fault diagnosis method based on multi-source information reinforcement learning. The method comprises the steps of carrying out the state monitoring of a steam generator, a water supply pump and a water supply valve in a steam generator water supply system, and obtaining a plurality of pieces of state information; clustering the associated state information based on the information source type to form a feature information cluster; establishing a feature matrix based on feature information cluster integration, and extracting a state feature vector from the feature matrix; and inputting the state feature vector into a data analysis model system to diagnose the fault type, and evaluating, regulating and controlling the state parameter of the corresponding component. Aiming at multi-source information fault diagnosis of a steam generator water supply system, technologies of signal processing, mutual information association, nonlinear dimension reduction, deep reinforcement learning and the like are adopted, so that the problems of feature extraction, feature fusion, feature integration, feature learning and the like of multi-source information are solved; and a new technical scheme is provided for autonomous learning and state recognition and diagnosis of multi-source information of the steam generator water supply system.
Owner:WUHAN SECOND SHIP DESIGN & RES INST

Water level control method of steam generator in nuclear power system based on sarsa

The present invention relates to the technical field of nuclear power system control, and provides a method for controlling the water level of a steam generator in a nuclear power system based on SARSA, including: Step 1: Designing a SARSA controller for controlling the water level of a steam generator in a nuclear power system: selecting the water level Part of the influencing factors and part of the control quantity are respectively used as the input and output of the SARSA controller, and the control objectives are set; Step 2: Separate interval divisions for the input and output parameters of the SARSA controller, and design reward rules; Step 3: Train the SARSA controller update the Q table; Step 4: If the training result does not meet the control target, adjust the parameters of the SARSA controller according to the training result, and go to Step 3; otherwise, save the Q table. The invention can simultaneously realize rapid, accurate and stable control of the water level of the steam generator, and realize that the water level controller of the steam generator has the ability of self-learning under variable working conditions and system characteristic decline.
Owner:空天推进(苏州)航空航天科技有限公司

Method, device and medium for energy-saving shutdown of base station

The present disclosure relates to a method, device and medium for energy-saving shutdown of a base station. This method combines the characteristics of the 5G network, and uses the characteristic data and real-time status data of the base station as the input of the shutdown decision model, including service scene characteristics, tidal characteristics and neighbor relationship characteristics, and uses the shutdown decision model to determine the shutdown decision model for the base station. Shutdown strategy, including shutdown mode, execution time period and trigger threshold, for the base station to execute. Use the feedback data to evaluate the impact. If the shutdown strategy is effective, add the shutdown strategy to the effective energy-saving case database, and perform multi-label regression model autonomous training based on the effective energy-saving case database to update the shutdown decision model. Form a closed loop.
Owner:CHINA TELECOM CORP LTD

Edge AI-based voice posture sensing electric toothbrush system

PendingCN113284483AIt has the function of independent closed-loop learningSelf-directed closed-loop learning abilityBrushesCharacter and pattern recognitionPower toothbrushSpeech output
E The invention discloses an Edge AI-based voice posture sensing electric toothbrush system, which comprises an MCU provided with an edge calculation training model, and comprises a voice acquisition mechanism, an attitude sensing mechanism, a voice output mechanism, a motor control mechanism, a networking mechanism and an indicator light mechanism which are simultaneously connected with the MCU. The MCU is also connected with a crystal oscillator for providing an accurate clock beat and a Flash storage chip. The edge calculation training model is provided with a decision model based on edge calculation training and a voice recognition model and a posture perception recognition model which are connected with the decision model at the same time, the voice recognition model is connected with an input voice processing module, and the posture perception recognition model is connected with a posture data processing module. The decision-making model is connected with a networking processing module, a state indicating lamp processing module, an indicating voice output processing module and a posture adjusting data output processing module at the same time.
Owner:深圳市回映电子科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products