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

531results about How to "Improve fitness" patented technology

Generative adversarial network-based pixel-level portrait cutout method

The invention discloses a generative adversarial network-based pixel-level portrait cutout method and solves the problem that massive data sets with huge making costs are needed to train and optimizea network in the field of machine cutout. The method comprises the steps of presetting a generative network and a judgment network of an adversarial learning mode, wherein the generative network is adeep neural network with a jump connection; inputting a real image containing a portrait to the generative network for outputting a person and scene segmentation image; inputting first and second image pairs to the judgment network for outputting a judgment probability, and determining loss functions of the generative network and the judgment network; according to minimization of the values of theloss functions of the two networks, adjusting configuration parameters of the two networks to finish training of the generative network; and inputting a test image to the trained generative network for generating the person and scene segmentation image, randomizing the generated image, and finally inputting a probability matrix to a conditional random field for further optimization. According tothe method, a training image quantity is reduced in batches; and the efficiency and the segmentation precision are improved.
Owner:XIDIAN UNIV

Automatically adjusted electric vehicle staged automatic emergency braking control system

The invention relates to an automatically adjusted electric vehicle staged automatic emergency braking control system. The automatically adjusted electric vehicle staged automatic emergency braking control system comprises vehicle-mounted distance measuring and speed measuring sensor equipment, a staged pre-warning control system, a safety distance calculation model, a vehicle reverse longitudinaldynamic calculation model, a hydraulic braking force and regenerative braking force distribution and calculation module, a hydraulic braking system reverse model, an ESC and Booster active pressurizing hydraulic pressure distribution module and a pavement information estimation module; potential safety hazards such as rear-end collision of a rear vehicle caused by great sudden deceleration of a vehicle are reduced while the comfort during triggering an AEB system is improved; by estimating information, such as gradients and adhesion coefficients, of a current running road and a front road ofthe vehicle and controlling online adjusting of parameters, the adapting degree of the AEB system to different pavement statuses is enhanced, the braking energy is sufficiently recycled, the endurancemileage is increased, and the advantages of a vehicle body stability control system ESC and an electronic mechanical braking booster in active pressurizing are sufficiently developed.
Owner:JILIN UNIV

Web service recommendation method based on socializing network platform

The invention relates to a Web service recommendation method based on a socializing network platform. Web services are classified into different service classes, basic information of users in the socializing network is obtained, and partials users are taken as training users and service classes which attract the users are obtained, so that a user-service matrix is set up based on the service classes. User clustering is carried out according to the user-service matrix, so that different interests of the users are obtained. In terms of new users who need Web service recommendation in the socializing network, recommendation is carried out according to the interests of the users. According to the Web service recommendation method based on the socializing network platform, theme extracting is carried out on the Web services, and the Web services are mapped to be the Web service classes, so that dimensionality reduction is carried out on the user-service matrix. Basic information and interest labels of the users are extracted from the socializing network and label information and semantic information are fully utilized for solving the synonymy problem of words. A clustering result is obtained by analyzing historical access records of existing users and Web service recommendation is carried out on the new users.
Owner:江苏友邦软件有限公司

Method and device for route optimization of logistics delivery vehicle

The invention discloses a method and a device for route optimization of logistics delivery vehicle, and belongs to the technical field of logistics. The method comprises the following steps of: initializing a congestion matrix alpha and a distance matrix D, generating a delivery route weight matrix omega=alpha D, and initializing a population module N<ZQ>; selecting a population size N<X>, a maximum number of generations N<G>, a crossing-over rate beta, a mutation rate gamma and a number of generations n=0, generating an initial route r1 through a greedy algorithm, and performing mutation operation on the initial route r1 to generate N<ZQ>-1 new routes; calculating fitness A<n> of each route of a first generation population formed by the initial route and the new routes, selecting N<X> routes with the highest fitness from the current population by adopting selection operators, and performing crossover and mutation operations on the N<X> routes to generate a population of next generation; updating n=n+1, when n=N<G>, calculating the fitness A<n> of all the routes in the latest population, and selecting the delivery route with the highest fitness in the current population as the optimal route. According to the invention, when the logistics delivery vehicle delivers goods, the delivery time can be as less as possible, and the delivery route can be as short as possible.
Owner:余意 +3

Distributed formation method of unmanned aerial vehicle cluster based on reinforcement learning

The invention discloses a distributed formation method of an unmanned aerial vehicle cluster based on reinforcement learning. The distributed formation method comprises the steps that step (1), a formation target state function and a simulation model of environmental uncertainty factors are obtained, and an unmanned aerial vehicle formation simulation model is established; step (2), under the interference of the environmental uncertainty factors, based on the unmanned aerial vehicle formation simulation model established in the step (1), a Q learning method is adopted to train the unmanned aerial vehicle cluster to update a flight strategy table; step (3), the value of the completion degree of the formation target state is calculated according to the obtained formation target state function, the obtained value of the completion degree of the formation target state is compared with a preset value of the formation target state, whether the formation target state is reached or not is judged according to the comparison results, if the formation target state is reached, a step (4) is performed, and if not, the step (2) is entered; and step (4), the updated flight strategy table is saved. According to the distributed formation method of the unmanned aerial vehicle cluster based on reinforcement learning, flight strategy parameters with adaptability are provided for the cluster, and the stability and robustness of the unmanned aerial vehicle cluster formation are guaranteed.
Owner:XIDIAN UNIV

Intrusion detection method fused with improved intelligent bee colony algorithm and BP neural network

The invention discloses an intrusion detection method fused with an improved intelligent bee colony algorithm and a BP neural network. The intrusion detection method includes the following steps: collecting network data packets and preprocessing the network data packets to take the preprocessed network data packets as training data of an intrusion detection model; designing a multi-layer neural network model, and setting activation functions for neurons of a hidden layer and an output layer; pre-training the neural network model by using the improved intelligent bee colony algorithm, and outputting the optimal honey source position vector; according to the optimal honey source position vector, setting the initial weight and threshold of the neural network model; designing a back propagation algorithm and using the intrusion detection data to train the neural network to obtain a neural network intrusion detection model; and designing a network intrusion detection software module and deploying the network intrusion detection software module in a network environment to detect the network data traffic in real time, and generating an alarm for the detected abnormal network traffic. Theintrusion detection method fused with the improved intelligent bee colony algorithm and the BP neural network adopts the improved intelligent bee colony algorithm to optimize the BP neural network, and can improve the training speed of the neural network model and the accuracy for intrusion detection.
Owner:SHANGHAI MARITIME UNIVERSITY

A raw material pretreatment method for improving the dissolution rate of active ingredients of medicinal plants

The invention discloses a raw material gas explosion pretreatment method for improving the dissolution rate of active ingredients of medicinal plants. The method comprises the following steps of: putting materials preimpregnated in water or aqueous solution into a high-pressure container, sealing, introducing a medium, keeping the temperature between 40 and 250 DEG C for 0.5 to 30 minutes, releasing pressure quickly, and discharging the materials in the high-pressure container to a normal-pressure container to obtain materials which are subjected to gas explosion pretreatment, wherein the medium is one or a combination of two of air, carbon dioxide, ozone, liquid nitrogen, nitrogen, water vapor and supercritical CO2. On the basis of the bioactivity of the active ingredients of the medicinal plants, various medium combination gas explosion modes in the low-temperature and high-pressure state are adopted selectively to ensure the treatment strength of the materials, and the dissolution rate of the active ingredients is improved on the premise that the active ingredients are not damaged to improve the utilization rate of raw materials; and the method is quick, simple, economic, high in adaptation degree and wide in selective range, and pollution is avoided, so the method is a significant path for realizing high-value utilization of the medicinal plants.
Owner:INST OF PROCESS ENG CHINESE ACAD OF SCI

Portable water quality monitor

InactiveCN102331485AImprove the situation that the number is limitedSimple designTransmission systemsTesting waterGeneral Packet Radio ServiceMicrocontroller
The invention provides a portable water quality monitor, which comprises an upper computer and a lower computer, wherein the upper computer and the lower computer are connected through a communication middleware or a wireless communication part, and the upper computer comprises an upper computer human-computer interface part; the lower computer comprises a singlechip, a data acquisition part, an expanded SDRAM (synchronous dynamic random access memory) part and a lower computer human-computer interface part, the data acquisition part is provided with an SDI (standard data interface)-12 sensor connected with the singlechip through a flexible SDI-12 interface bus, and the SDI-12 interface bus is connected with at least one SDI-12 sensor through a successive type interface; the wireless communication part comprises a GPRS (General Packet Radio Service) wireless module; and a power source part comprises a selfcontained power source, a solar energy power source and a power source management module used for switching the power source into the selfcontained power source when voltage provided by the solar energy power source is not sufficient to maintain the work of the water quality monitor. In the invention, the number of sensors accessed into the portable water quality monitor is not limited, the data transmission is simple, and the data storage capacity is large.
Owner:EAST CHINA UNIV OF SCI & TECH +1

Hyperspectral remote sensing image target detecting method based on variable end members

The invention discloses a hyperspectral remote sensing image target detecting method based on variable end members, comprising the following steps of: selecting a remote sensing image to be processed by target detection; acquiring prior information required for detection, wherein the prior information comprises spectral information of target end members and spectral information of background end members; traversing the remote sensing image to be detected by utilizing a cross correlation matching technique to determine the types of background end members in each pixel in the remote sensing image to be detected; carrying out spectral decomposition on the remote sensing image to be detected in a completely restricted least square way to acquire the component information of target end members and various background end members in each pixel in the remote sensing image to be detected; establishing a detector based on the GLRT (Generalized Likelihood Ratio Test); and traversing the remote sensing image to be detected by adopting the detector to acquire the detection function value of each pixel in the remote sensing image to be detected, thereby judging whether targets exist in each pixel in the remote sensing image to be detected or not. The method of the invention has the characteristics of strong structuration, high adaptability, self-organization and self-learning.
Owner:WUHAN UNIV

Automatic network cable distributing device and distributing method

The invention provides an automatic network cable distributing device which comprises a casing, cables, a baseplate, a cable outlet pillar, inserting and pulling mechanisms, a driving part, a cable plug, a sensor and a controller. The controller starts the driving part according to pre-jumping information and information transmitted by the sensor, and the driving part drives the inserting and pulling mechanisms on two sides of an inserting port to move, so that the cables can start from an initial position and accurately reach the inserting port needing jumping and at the target position according to the positioning sensor. Specifically, the inserting and pulling mechanisms are clamped and pull the cable plug of the cables in pre-jumping out, the cables get away from the initial position, are sequentially wound according to the cable winding rules and move towards the target position, and after reaching the target position, the cables are inserted into a socket at the target position, and cable distribution is finished. The automatic network cable distributing device is compact in structure, simple in dimension, easy to install and maintain, fast to start, short in time for finishing the process, less in components and low in cost and simple in control principle compared with the existing like products.
Owner:北京瑞祺皓迪技术股份有限公司

Automatic pilot and genetic algorithm-based method for monitoring hovering range of unmanned aerial vehicle

The invention relates to an automatic pilot and genetic algorithm-based method for monitoring the hovering range of an unmanned aerial vehicle. According to the method, when the unmanned aerial vehicle malfunctions or rolls over, an automatic control instruction is generated for an automatic pilot; and the automatic pilot and an genetic algorithm can be combined together so as to perform unmanned aerial vehicle flight hovering control; when it is monitored that sharp pitching or rollover of the unmanned aerial vehicle is not brought about by a flight instruction of a task, the automatic pilot immediately adopts a flight attitude adjustment parameter as a local initialization population value; the genetic algorithm performs iterative optimization on the adjustment quantity of an adaptive value according to the requirements of adaptability conditions in the initialization value; an adjustment solution of a control instruction when the flight attitude of the unmanned aerial vehicle tends to be stable can be obtained through iterative selection; and therefore, stable control on the attitude of the unmanned aerial vehicle can be realized, and the unmanned aerial vehicle can hover or land in a certain space area range.
Owner:BEIJING CASRS INFORMATION TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products