4162results about How to "Improve rationality" patented technology

The invention discloses a method for estimating the charge state of a power cell, which comprises the following steps that: 1. a continuous state space model capable of expressing the relationship of each element in an expression circuit is obtained by an equivalent circuit model of a power cell; the relationship between the open-circuit voltage and the cell charge state is obtained through the standing experiment of the power cell, and the cell charge state is introduced into the continuous state space model of the power cell as a state variable; the power battery model under the noise environment is obtained by combining the noise information; and finally the continuous state space model is subjected to linearization and discretization so as to obtain a linear discrete state space model; 2. a relationship curve of the open-circuit voltage and the cell charge state is obtained through the standing experiment of the power cell, and a parameter k and a parameter d in the power cell model are obtained approximately; 3. parameters of the power cell model are obtained through the identification of current and voltage data collected by a data collecting system; and 4. the power cell charge state is estimated by utilizing a moving horizon estimation method based on the determined power cell model.

The invention discloses a method for identifying abnormal motion in a video. The method comprises the steps that an area of which the pixels have certain change is extracted from a video sequence through a foreground extraction module; a movement target detected by the foreground extraction module is judged through a pedestrian judgment module, and a rectangular box judged to be a pedestrian is kept and transmitted to a target tracking module; each identified target in the scene is continuously tracked in a multi-target way through the target tracking module; the posture of each target tracked by the target tracking module within each frame of time is judged through an abnormal behavior trigger module and behavior analysis is performed on the abnormal behavior; and the abnormal behavior is processed and computed into a video segment through a behavior identification module, and the abnormal behavior is alarmed and the behavior type of the abnormal behavior is identified. The invention also discloses a system for identifying the abnormal motion in the video. The problem that abnormal behavior (falling down) detection of massive video data cannot be performed in real time can be solved and the detection accuracy can be greatly enhanced.

The invention discloses an event atlas construction system and method based on multi-dimensional feature fusion and dependency syntax. The event graph construction method based on multi-dimensional feature fusion and dependency syntax is realized through joint learning of event extraction, event correction and alignment based on multi-dimensional feature fusion, relationship extraction based on enhanced structured events, causal relationship extraction based on dependency syntax and graph attention network and an event graph generation module. According to the event graph construction method and device, the event graph is constructed through the quintuple information of the enhanced structured events and the relations between the events in four dimensions, and the defects that in the priorart, event representation is simple and depends on an NLP tool, the event relation is single, and the influence of the relations between the events on event graph construction is not considered at the same time are overcome. According to the event atlas construction method provided by the invention, the relationships among the events in four dimensions can be randomly combined according to different downstream tasks, and the structural characteristics of the event atlas are learned to be associated with potential knowledge, so that downstream application is assisted.

The invention discloses a staged multi-path model algorithm of urban rail transit network passenger flow distribution. The algorithm comprises the following steps of: dividing definite factors and indefinite factors affecting the outgoing of passengers into two stages to calculate and correct the passenger flow distribution proportion of a path, determining the initial distribution proportion of the passenger flow on the path by using the comprehensive outgoing impedance of time unit as a main factor, and correcting the proportion by considering the transfer times and the crowdedness degree on the basis of time window constraints according to a generated full origin-destination (OD) passenger flow distribution path set. The model and the algorithm provided by the invention have higher reasonability and practicability and can effectively serve the passenger flow management of the urban rail transit, meet the service requirement of clearing automatic fare collection (AFC) passenger flow and fare, and provide auxiliary decision proof for operation organization optimization and coordination.

The invention relates to a method and a system for space and time analysis of urban road traffic states. On the basis of a space and time autocorrelation theory, a basic theory of mathematical statistics is applied, and according to the real-time traffic flow data of a road network, the space and time evolution of the traffic states and the generation, development and dissipation process of congestion are analyzed. The distribution features of different road states in space and time are subjected to statistical analysis by the method, the space and time rule of the evolution of the road traffic states of the urban road network is analyzed, the congestion rush hour in time and the bottleneck road in space are identified and judged, and the formulation and implementation of traffic measures are further supported.

The invention discloses an active learning based data automatic marking method, and belongs to the field of active learning. The data automatic marking method comprises the steps of 101, processing marked data and unmarked data; 102, classifying the unmarked data by using a plurality of different classifiers; 103, selecting data with low difference entropy; 104, performing manual marking on the data with low difference entropy; and 105, performing self-checking on a manual marking result. The invention provides a data automatic marking system with a self-checking function through combining an active learning method in allusion a problem of how to ensure the accuracy of manually marked data as far as possible while reducing the volume of the manually marked data, thereby achieving the purposes of reducing the workload and improving the accuracy of the manually marked data.

The invention discloses a defrosting control method. The defrosting control method comprises the following steps: controlling an air conditioner in a steady state, and remaining to operate for some preset time in the steady state; in real time detecting the operation current of an outdoor fan; judging whether the change rate of the operation current of the outdoor fan is not more than a preset value under a preset period or not, if so, controlling the air conditioner unit in a defrosting state. The defrosting process is controlled according to the current change rate of the outdoor fan of the air conditioner, in comparison with the prior art, the defrosting process is controlled according to the temperature of the outdoor tube of the air conditioner and the current value of the outdoor fan, the operation is relatively simple, the defrosting process is relatively reasonable and the defrosting effect is relatively better. The invention further disclose a defrosting control device.

The invention discloses a dialogue generation model training method, which comprises the following steps: obtaining a dialogue data set, the dialogue data in the dialogue data set comprising questionsand annotation replies corresponding to the questions; based on the questions in the dialogue data set, obtaining the coded representation of the question through an encoder in a constructed dialoguegeneration model; fusing the coded representation of the question and knowledge information of transfer learning of the question from a knowledge base question and answer model through a decoder in the dialogue generation model to obtain a prediction reply corresponding to the question output by the dialogue generation model; and determining a loss function based on the prediction reply and the annotation reply corresponding to the question, and adjusting parameters of the dialogue generation model through the loss function until the loss function of the dialogue generation model is convergent. According to the method, the knowledge information can be better fused by the model, so that the dialogue generation accuracy and reasonability are improved. The invention further discloses a dialogue generation method and device, equipment and a medium.

The invention discloses an automatic driving overtaking decision-making method based on reinforcement learning under opposite double lanes. The method comprises the following steps of collecting the traffic state of an automatic driving vehicle through a sensor; inputting the collected traffic state into a trained decision model; enabling the decision model to select a corresponding driving actioninstruction from the action space according to the input information and output the driving action instruction, and driving the vehicle automatically to form a new traffic state after the driving action; calculating a reward value of the driving action through a reward function, and storing the original traffic state, the driving action, the reward value and the new traffic state into an experience playback pool as transfer samples; calculating a loss function value of the decision model, and optimizing parameters of the decision model according to the transfer sample and the loss function value; and repeating the steps until the automatic driving is finished. The safety and comfort of the overtaking decision-making process of the automatic driving vehicle are ensured, and the humanization and robustness of decision-making are improved through a reinforcement learning decision-making method.

The invention discloses an intelligent inquisition method, an intelligent inquisition system, computer equipment and a storage medium. The intelligent inquisition method includes the following steps that: disease prior probability prediction is performed according to the main complaint content of a target object, disease prior probabilities for the target object are obtained; at least one candidate disease is determined based on the order of the disease prior probabilities; and at least one agent strategy correspondingly associated with the at least one candidate disease is determined from a pre-established multi-agent model based on reinforcement learning, wherein each agent strategy is used to indicate questions which are required to be raised for the output of next interaction under a current symptomatic state; based on each agent strategy, interactive consultation is conducted according to the interactive selection of the questions which are required to be raised for the output ofthe next interaction for the target object. With the intelligent inquisition method, the intelligent inquisition system, the computer equipment and the storage medium of the invention adopted, the rationality of inquisition logic for related diseases and user experience can be optimized with the accuracy of inquisition results ensured.

The invention relates to an establishment method for hydrocarbon source rock hydrocarbon production rate charts in petroleum resource assessment. The establishment method includes the steps of 1), collecting experimental samples and determining thermal simulation experiment data through experiments; 2), collecting materials; 3), calibrating kerogen generated oil, generated gas and oil-cracking gas kinetic parameters; 4), establishing sedimentary burial history and thermal-history models of hydrocarbon source rocks on a target stratum of a research area; 5), establishing constraint condition of hydrocarbon source rock hydrocarbon production rate; 6), establishing kerogen generated oil, kerogen generated gas, oil generated gas, clean and total gas conversion profile charts on the target stratum of the research area; 7), checking whether dynamics geological extrapolation results meet requirements or not; 8), putting forwards part of hydrocarbon expulsion mode adjustment coefficients, evaluating hydrocarbon production rate curves of the partial hydrocarbon expulsion model, and establishing the hydrocarbon source rock hydrocarbon production rate charts including three modes, namely the mode of complete hydrocarbon expulsion, the mode of partial hydrocarbon expulsion and the mode of no hydrocarbon expulsion. The shortcoming in an original calculating method and experiment simulated data of the hydrocarbon source rock hydrocarbon production rate is overcome, and the hydrocarbon source rock hydrocarbon production rate at the target stratum can be accurately and reasonably determined according to oil field exploration practice by oil field workers.

A technology for integrally casing large bearing steel roller features that the computer simulation technique is used to design metal mould, the sand covered iron mould and insulating raiser head are used for sequentially solidifying, and the new pouring system and the rotary filling of tangent internal sprue are used for preventing air bubble and impurities, resulting in high quality of product.

The invention discloses a method and a device for realizing load balance among neighboring cells in order to solve the problems in the prior art such as low efficiency of the realization of the load balance among the neighboring cells and low after-balancing mobile communication network quality. The method comprises: determining to-be-regulated cells requiring load balance currently according to the load ratio of cells; according to the performance information of the neighboring cells of the to-be-regulated cells, selecting target cells to participate in load balance with the to-be-regulated cells from the neighboring cells; according to the performance information of users in the to-be-regulated cells, selecting target users to participate in load balance from the users in the to-be-regulated cells; and adjusting the target users to the target cells. According to the technical scheme of the invention, the effectiveness of the realization of load balance among neighboring cells is improved and the quality of a mobile communication network is improved.

The invention discloses a method and device for controlling an electronic expansion valve. The method for controlling the electronic expansion valve includes the steps of a, periodically obtaining the current outdoor environment temperature and the current frequency of a compressor, obtaining a first adjusting range of the electronic expansion valve according to the adjustment range, corresponding to the preset outdoor environment temperature, of the electronic expansion valve, obtaining a second adjusting range of the electronic expansion valve according to the adjustment range, corresponding to the preset frequency of the compressor, of the electronic expansion valve, b, obtaining the intersection of the first adjusting range of the electronic expansion valve and the second adjusting range of the electronic expansion valve, taking the intersection as a final adjusting range of the electronic expansion valve, c, obtaining the variation trend of exhaust gas temperature, and correspondingly adjusting the electronic expansion valve within the final adjusting range of the electronic expansion valve according to the variation trend of exhaust gas temperature. According to the method and device for controlling the electronic expansion valve, the electronic expansion valve is adjusted fast and reasonably under the control of the outdoor environment temperature factor adjusting range and the compressor frequency factor adjusting range and on the basis that the electronic expansion valve is controlled by the exhaust gas temperature.

A composite energy system of solar energy, ground source heat pump and chilled water storage comprises an air conditioner, a domestic hot water supply device, one or more device (hereinafter referred to as host) of a solar energy heat-collecting device, a heat storage water tank unit, a cold storage water tank unit and a ground source heat pump system, a ground heat exchanger, an auxiliary heating device, a circulating water pump, a first electric three-way valve, a second electric three-way valve, a first electric two-way valve and a second electric two-way valve, wherein the circulating water pump is communicated with the air conditioner, and the auxiliary heating device is communicated with the domestic hot water supply device. The invention fully utilizes renewable energy sources such as solar energy, geothermal energy and biomass energy to masterly combine the ground source heat pump air-conditioning system with the solar energy hot water system and the chilled water storage system and construct a comprehensive energy system combined by the domestic hot water system, the ground source heat pump, the chilled water storage air-conditioning system and the like; and the invention is completely combined with the entire building so as to greatest reduce the floor space and the investment and increase the energy efficiency and economic performance of the system.

The invention discloses an explicit congestion control method based on bandwidth estimation in a high-bandwidth delay network, which comprises the following steps of: calculating the current available bandwidth of a network by a TCP (Transmission Control Protocol) sender through observing an ACK (Acknowledgement Character) receiving interval, loading a router by a packet header ECN (Explicit Congestion Notification) 2bits information feedback network, carrying out network state judgment by the sender through explicit feedback information and estimated available bandwidth values, and adjusting a congestion mechanism according to a judgment result. The congestion is judged by combining end-to-end feedback and router explicit feedback, i.e. finer congestion information is given by using the bandwidth estimation values based on the explicit feedback of coarseness, thus an algorithm ensures high congestion feedback precision while reducing the explicit feedback overhead and improving the protocol popularization, and improves the efficiency and the convergence rate of the protocol.

The invention provides a method and a device for improving the accuracy and energy-saving efficiency of an 'on-off time-area process' heat allocation system. The device comprises total heat supply metering devices arranged in building heat supply pipes and room temperature controllers and electrical heat balance flow regulating devices which are installed by heat supply users respectively, wherein the room temperature controllers are arranged at reference positions in heating areas of the users; and the electrical heat balance flow regulating devices comprises on-off control valves which are arranged in house branch heat supply pipes of the users. The room temperature controllers monitor indoor temperatures, and the electrical heat balance flow regulating devices control the flows in the heat supply pipes and the opening and closing of the heat supply pipes, and calculate and store the total opening time length of on-off control valves.

The invention discloses an enterprise power consumption load prediction method based on a k-means clustering radial basis function (RBF) neural network. The method includes steps: historical load data acquisition, meteorological data acquisition, date discrimination, neural network prediction, error calculation and correction, load curve drawing, and prediction data export. A prediction result is obtained by employing the neural network via the historical load data and meteorological factor input quantity, and correction is realized via an error correction module. Based on the requirement control technology of load prediction, with the combination of an industrial enterprise production plan and the condition of power consumption load usage, the system performs requirement control via a built-in requirement curve node determination method at a control point before the load prediction value reaches the maximum requirement, whether unnecessary loads are removed is determined, the current most appropriate energy-saving scheme is automatically selected, the maximum requirement is controlled in advance, over-load operation and even tripping operation can be effectively avoided, and safety and energy-saving production is guaranteed.

The invention puts forward a synonym expansion method and device for search information. The method comprises the following steps that: carrying out word segmentation processing on the search information to obtain at least one segmented word of the search information; obtaining the candidate synonym set of the segmented word, wherein the candidate synonym set comprises at least one synonym of thesegmented word; forming a synonym pair which contains the synonym and the segmented word by aiming at each synonym; carrying out feature extraction on the synonym pair to obtain a synonym pair featureset; according to the feature set, predicting the synonym pair to obtain a target probability that the synonym pair is predicted as reasonable replacement; and if the target probability exceeds a preset threshold value, forming a synonym expansion item by the segmented word and the synonym, and searching and obtaining a search result on the basis of the synonym expansion item. Through the method,synonym replacement rationality and accuracy can be improved, the recall rate and the accuracy of a search result are improved, and therefore, the technical problems in the prior art that synonym replacement is inaccurate and a search result recall rate is poor can be solved.

The invention discloses a method for generating an urban bus route network based on segmented splicing. The method comprises the following steps that first, urban bus stations are used as network nodes, and a user balancing distribution method is used for acquiring the passage point traffic of all the bus stations; a bus center point determining method is used for acquiring a bus center node set, and a bus network route section set of bus single center nodes among bus center nodes is generated; accordingly, an initial line network generation method is used for setting the departure frequency of all bus routes, and an initial bus line network planning scheme is obtained; then, a simulated annealing algorithm is used for optimization; a novel bus line network planning scheme is formed through a line network adjusting method and a frequency adjusting method every time, and a strategy balancing distribution method based on effective frequencies is used for simulation and evaluation until an optimized urban bus line network is acquired. The method for generating the urban bus route network based on segmented splicing has the advantages that the steps are simplified, the practicability is high, standardized implementation can be achieved, and the planning routes are reasonable and scientific.