52results about How to "Reduce iteration time" patented technology

The invention discloses a robot positioning method. A to-be-positioned robot is accessed to a wireless local area network on the basis of WiFi. The robot positioning method comprises steps that estimated present location positions of the robot are determined based on WiFi fingerprint positioning; a map of the estimated positions is called, the constraint estimated positions of the map are marked, and the precise position of the robot is acquired by employing a particle filtering algorithm. The robot positioning method is advantaged in that positioning precision is relatively high, and the required iteration time is relatively short.

The invention discloses a post-disaster repair decision-making method and device for power distribution network. The method comprises the steps of collecting fault information distribution transformerlevel disaster damage information,first-aid repair teams and material transport vehicle information under disaster conditions; on the premise of meeting the distribution network operation constraintand the path planning constraint,using distribution network fault point selection and first-aid repair teams and material scheduling as decision variables,establishing a power distribution network frame reconstruction and material personnel allocation auxiliary decision-making model,wherein the model comprises a lower tie line action scheme determination model taking the maximum recovery load as atarget and an upper construction team and material scheduling scheme determination model taking the minimum total power loss in the rush-repair process as a target; and solving the dispatching auxiliary decision-making model by adopting an upper-lower layer iterative solving method,and finally determining a post-disaster repair decision-making scheme of the power distribution network. According to the method,the problem that the target of minimum economic loss cannot be achieved due to the cooperative action of a construction team and a material vehicle is fully considered,and decomposition codes of all parts are simplified.

The invention discloses an improved brute-force search method based train vehicle energy-saving operation method. The method includes steps of setting train operation line data and train parameters and establishing a train operation power consumption calculation model; taking train traction energy consumption and interval travel time as optimization targets, establishing a train operation energy-saving optimization model; processing an objective function and constraint conditions through a goal constraint method and converting a multi-target problem to a single-target problem; processing the constraint conditions, narrowing a solution search range and processing the converted single-target problem through a brute-force search method; selecting a train operation strategy with punctuality and the highest energy conversation rate according to the set train interval travel time and the interval distance, and obtaining a corresponding speed curve and energy consumption curve. According to the invention, regular energy conservation optimization precision of a train is improved effectively. Besides, the optimization speed is high and train energy consumption is reduced.

The invention discloses a Cam-shift-based prediction tracking method. A Kalman filtering and Cam-shift algorithm combination method is proposed; and a linear prediction and Cam-shift algorithm combination-based target prediction tracking method is proposed, namely, a linear prediction method replaces Kalman filtering to finish prediction estimation, and a prediction estimation result is substituted into a Cam-shift algorithm to perform tracking. In order to well finish position prediction tracking of a moving target in a complex background with interference of factors such as shielding and the like, the Kalman filtering and Cam-shift algorithm combination method is researched. The Kalman filtering can finish position prediction of the moving target more accurately; and the tracking can be well finished by combining the Kalman filtering with the Cam-shift algorithm. Based on this, the linear prediction and Cam-shift algorithm combination-based target prediction tracking method is proposed, namely, the linear prediction method replaces the Kalman filtering to finish the prediction estimation, and the prediction estimation result is substituted into the Cam-shift algorithm to perform the tracking. An experiment shows that the method can not only ensure the accuracy of tracking in a shielding process but also shorten the iterative time, and can better meet the requirement of timeliness.

The invention discloses a detection system based on a QT platform and a use method thereof. The invention relates to a detection system based on a QT platform, in particular to a detection system based on the QT platform and used for detecting an airplane. According to the invention, the system and the method are convenient to operate, clear in logic and high in modification speed, an operator ofthe detection system can directly modify the detection system conveniently and efficiently according to micro modification of control / output data in a subsystem of the detection system, so that a designer of the detection system does not need to repeatedly update the detection system for multiple times. The system comprises an interface layer, a data layer and a communication layer, the communication layer comprises a packing module, an unpacking module and a communication module and is used for completing encoding and decoding of communication data frames and setting of necessary communication parameters, so that the detection system can communicate with an underlying system. According to the invention, the research and development and test cost of a detection system is saved, the use difficulty is reduced, and the use effect is excellent.

The invention discloses a pit detection method based on improved YOLOv3, and the method comprises the steps: S1, collecting a pit image through a visual collection system, and obtaining a pit data set after preprocessing, wherein the pit data set comprises a preprocessed pit image; S2, constructing an improved YOLOv3 pit detection network model; S3, inputting a training data set of the pit data set into the improved YOLOv3 pothole detection network model for training, and when the improved loss function approaches zero, obtaining a parameter optimal solution of the improved YOLOv3 pothole detection network model; and S4, inputting the pit data set into the improved YOLOv3 pit detection network model into which the parameter optimal solution is substituted to obtain a pit detection result. According to the invention, the problems that the real-time performance of pit detection needs to be guaranteed, and the accuracy needs to be further improved are solved.

The invention discloses a hardware-in-the-loop simulation test system for a vehicle-mounted computing unit of an automatic driving vehicle, which can shorten the debugging and iteration time of a real vehicle vehicle-mounted computing unit and solve the problems of low efficiency and high consumption of direct real vehicle test. The simulation test system can test a real vehicle-mounted computing unit suitable for a structured road driving environment, simulate all hardware equipment in an automatic driving system including sensor hardware, the vehicle-mounted computing unit, a drive-by-wire system and the like, and provide a whole set of automatic driving test environment for developers; manual / automatic driving switching and actual application effect evaluation of the vehicle-mounted calculation unit can be realized, the response capability of the real vehicle-mounted calculation unit to various conditions can be comprehensively tested, the problems of the real vehicle-mounted calculation unit can be positioned, and the development time of a vehicle automatic driving system is saved; and the feasibility and the stability of a data transmission format and a communication mode used in a real vehicle automatic driving system can be tested.

Using a vector-based emulation technique, a hardware-based prototyping system reduces time-consuming recompilation and reduces the iteration time for a verification run. The vector-based emulation technique takes advantage of information derived from user-defined probe points, automatically generated probe points and low-latency snapshots. Using a bounded-cycle simulation technique, the hardware-based prototyping system can provide complete or partial simulation traces covering interested signals and can efficiently evaluates assertions. A user is therefore able to debug in a real system test and to identify causes of fault conditions interactively under a controlled vector debugging environment.

The invention relates to a decision level prediction method, a storage device and a storage medium, and the method comprises the steps: determining a storage unit where storage data is located under the condition that the storage data is read from a memory; acquiring the state information of the storage unit at the current moment; and determining a read judgment level of the storage unit at the current moment according to a corresponding relation between historical state information and a historical optimal judgment level and the state information of the storage unit at the current moment. According to the method, the current read decision level is predicted by querying the historical information of the storage particles and combining the state information of the storage unit, and the RBER of read data is effectively reduced, so that the iteration time and the corresponding power consumption of an ECC decoder are reduced, and the service life of the whole SSD equipment is correspondingly prolonged. Furthermore, according to the method, the physical characteristics of the read judgment level of the storage unit are obtained in an online learning mode, offline testing of particles is not needed, the testing workload is saved, and the adaptability is better.

The invention relates to the technical field of map construction, and relates to a construction method of an indoor 3D point cloud semantic map. The construction method comprises the following steps:1, acquiring an RGB-D image of an indoor environment by utilizing an RGB-D camera; 2, constructing a deep convolutional neural network Mask R-CNN which can be used for target detection and instance segmentation; 3, inputting the acquired RGB-D image into a network, and then performing point cloud processing on an output image; 4, fusing the image processed by the Mask R-CNN network and the point cloud with the constructed point cloud map of the indoor environment to obtain a semantic point cloud map; and 5, carrying out global optimization on the point cloud semantic map.

The invention discloses a novel cylindrical three-dimensional reconstruction method and system. The method comprises the following steps: acquiring dynamic scanning data of a line laser displacement sensor in a measuring mechanism on left and right journals of a wheel set axle, wherein the scanning data is three-dimensional point cloud data; performing initial calculation on the three-dimensionalpoint cloud data, and generating initial fitting parameters of the axle journal; and performing iterative optimization on the initial fitting parameter by using a particle swarm algorithm to obtain afinal fitting parameter. According to the method, during subsequent particle swarm algorithm iteration, because the precision of primary iteration is improved, the iteration time is reduced, and the problem that the particle swarm algorithm falls into local optimum is avoided, so that an optimal fitting result is obtained; the manufacturing precision and the assembling precision of the wheel set shaft neck are improved, and the running safety of a high-speed rail train and the riding comfort of passengers are also improved.

The invention discloses a distributed tool configuration construction generation method and system. The distributed tool configuration construction generation method comprises the steps: inputting a business package of each business system and metadata of the business package into a standard data management subsystem; dynamically loading the business package, analyzing the business package to obtain metadata of each service, and inputting the metadata of each service into a service management subsystem; receiving a tool construction request which is sent by the client and carries tool configuration information; querying metadata of a target service corresponding to the tool configuration information from the service management subsystem according to the tool construction request; and configuring the business query tool page according to the metadata of the target service, and generating a business query tool corresponding to the client. According to the scheme, personalized customization of the tools is achieved; user requirements are well met; each tool can be dynamically published in real time; the development time of the front end and the rear end is saved through configurationand automation; and the tool development efficiency is improved.

The invention provides a decision level prediction method, a storage medium and SSD (Solid State Disk) equipment. The method comprises the steps: when data reading is carried out, acquiring the use state information, the read data temperature and the write data temperature of a storage unit, wherein the read data temperature is the current temperature of the storage unit, and the write data temperature is the temperature of the storage unit when data is written; predicting a decision level of the storage unit according to the use state information; and determining a decision level compensation value according to the read data temperature and the write data temperature, and correcting the predicted decision level according to the decision level compensation value. According to the invention, the use state information of the storage unit in the temperature-crossing scene is calculated, and the optimal decision level is predicted based on the use state information, so an optimal judgment level is adopted for data reading during data reading, the RBER of read data caused by temperature-crossing reading and writing is effectively reduced, the iteration time and the corresponding power consumption of an ECC decoder are reduced, and the service life of the whole SSD equipment is prolonged.

The invention relates to a laser coherent array distributed phase control system and control method. The system comprises a laser coherent array output module and a distributed phase control module, wherein the distributed phase control module comprises a lenslet array, a plurality of secondary phase modulators, a plurality of optical fiber couplers, an optical fiber end cap, a plurality of photoelectric detectors, a primary phase control system and a secondary phase control system; a two-stage phase controller can effectively control wavefront information of array laser, and real-time control over the coherent combination effect of the array laser is achieved; and the optical fiber couplers perform pairwise coupling detection on the phase information of the array light beams without performing unified detection on the phase information of the whole array light beams, so that the phase control bandwidth of the system is improved. The laser coherent array distributed phase control system has the all-fiber feature, the pressure in the aspect of thermal management is reduced, and the compactness and operability of the system structure are improved.

The invention discloses a transformer fault diagnosis method based on CQFPA-WNN, and the method specifically comprises the steps: 1, collecting oil fault feature gas concentration data of an oil-immersed transformer, enabling the fault feature gas concentration data to serve as a total sample set, and dividing the total sample set into a training sample and a test sample; 2, performing normalization processing on the collected total sample set; 3, initializing a wavelet neural network and a cloud quantum flower pollination algorithm, and preparing for inputting wavelet neural network optimization parameters into a training sample; 4, optimizing wavelet neural network parameters by using the training sample and applying a flower pollination algorithm, and training the optimized wavelet neural network to obtain a diagnosis model; and 5, applying the test samples to a wavelet neural network transformer fault diagnosis model optimized based on a cloud operator flower pollination algorithm,and classifying the test samples to complete fault diagnosis. According to the method, the fault diagnosis speed and accuracy can be effectively improved.

A leaping iterative composition method of a complicated graphic and a storage medium having a computer program executing the same are described. First, an initiator and generators are set. After several times of iterating, a transitional object is formed. Then, a leaping recursion is performed based on this transitional object. When performing the leaping recursion, a generator of each of the iterations is designed by a single pattern converted from a structural result of a previous iteration added with a base object. Since the result of each of the iterations keeps the original structure, the structure is used as the input initiator. The initiator of the first iteration can be a feature of different iteration objects. Since the input structure of iterations are the same, the advantage of reducing computing resources and avoid system overload are reachable.

The invention discloses a rapid convergence laminated imaging device capable of synchronously collecting double defocus diffraction patterns, which comprises a light source, a focusing lens, a sampleand a beam splitter prism which are arranged in sequence, wherein light emitted by the light source passes through the focusing lens to form a focused light beam, after the focused light beam acts with the sample, the focused light beam is divided into a transmission part and a reflection part through the beam splitter prism, a first CCD and a second CCD are respectively placed at different positions of the transmission light path and the reflection light path away from the focus of the focused light beam, the first CCD is placed in front of the focus point, and the second CCD is placed behindthe focus point. According to the invention, the focusing lens, the beam splitter prism and the two CCDs are utilized to synchronously acquire a plurality of images with a plurality of defocusing amounts, the acquired image information is substituted into a light intensity transmission equation to obtain complex amplitude information of the CCD surface, and the information is utilized to construct a reasonable initial guessing value of an object to be measured, and the initial guessing value is substituted into a laminated imaging algorithm, so that the iteration time can be greatly shortened.

The invention provides a data iteration method and device. The method comprises the steps that when iteration of a data unit in a data file is conducted, meta information corresponding to the data unit is accessed, and a version number of the meta information is obtained; a name of the data unit and the version number of the meta information are combined, and based on the combination, a traversalquery is conducted on the data unit in the data file, wherein the version number of the data unit is included in a name field of the data unit in the data file. According to the data iteration method,a needed data unit can be rapidly found without needing to firstly conduct traversal query on data units with the highest ranking, data iteration time is saved, and rapid iteration can be conducted even after instant deleting.

The invention relates to the field of system performance monitoring, and discloses an application performance monitoring tool based on pinpoint so as to intelligently analyze performance data and automatically position a performance bottleneck. According to the invention, secondary development is carried out on the basis of a third-party open source tool pinpoint; data reported by a probe program are received through a collector program in a manner of injecting, for a Java application program, bytecode into a probe, affairs, application topographies, JVMs and other types of data are analyzed and stored, and finally a corresponding performance monitoring report is generated through the performance bottleneck analysis module, so that research personnel, test personnel and operation and maintenance personnel can know and solve system performance problems and bottlenecks in time. The method is suitable for application performance monitoring.

The embodiment of the invention discloses an online detection method for the surface shape of a reflector. On the basis of the principle of stripe reflection, a stripe image with ideal periodic changeis displayed through a display, a stripe image formed by the reflection of a reflector machined by an ultra-precision cutting machine is collected through an aperture camera, the surface shape of thereflector is judged and reconstructed according to the curvature of the stripe image, and thus, online detection of the surface shape of the reflector is completed. By using the online detection method for the surface shape of a reflector provided by the embodiment of the invention, a reflector with any smooth continuous surface shape can be detected without optical compensation device or precision position adjustment. The method has a greater dynamic range than the existing online detection method based on a laser interference detector, and takes shorter time and is less affected by the precision of the machine tool and the environmental conditions than the existing online detection method based on point-by-point ranging and scanning by a contact or non-contact sensor.

The invention provides a tunnel face supporting pressure design method. The tunnel face supporting pressure design method comprises the following steps: setting a target safety coefficient for maintaining short-term stability of a tunnel face and an initial value of supporting pressure according to the current engineering situation; calculating an actual safety coefficient according to the actualparameters of the tunnel structure and the strength reduction method, and continuously iterating until the calculated safety coefficient is the same as the target safety coefficient; in the iterationprocess, serving the ratio of the support pressure of each step to the load reduction coefficient as the support pressure applied by the next iteration step, performing strength reduction calculationto obtain the safety coefficient, and stopping the operation until the calculated current safety coefficient is equal to the target safety coefficient, wherein the corresponding support pressure is the target value at the moment. According to the method, the iteration time is greatly shortened and is generally more than ten hours, and the shortest iteration time is about two hours; and the problems that an existing method is low in efficiency, long in consumed time and not accurate enough are solved.

The invention discloses a two-dimensional static magnetic field parallel finite element acceleration method based on black box transmission line model. The method includes steps of 1, determining a to-be-solved variant and a solution domain; 2, establishing a rz rectangular coordinate system; 3, the governing equations and boundary conditions in the two-dimensional axisymmetric nonlinear static magnetic field are listed and a differential equation group is formed. 4, dividing that solution domain into networks; 5, calculating the coefficient matrix of the finite element; 6, the equivalent black box circuit model is established. 7, inserting a transmission line segment between that black box circuit model and the linear circuit; 8, carrying out transmission line method iteration; 9, repeating that iteration until the iteration result converges to a fixed error and the solution is end. The invention can solve the problem in parallel, and can be used for solving the complex finite elementnetwork model, thereby solving the problem of long solving time and low efficiency when Newton iterative method is used for solving the finite element non-linear problem.

A method and a system for establishing a metal interconnection layer capacitance prediction model are disclosed. The method for establishing the metal interconnection layer capacitance prediction model includes: extracting capacitance data of metal interconnect layer capacitors of different sizes by using a post simulation tool and establishing a relationship formula between capacitance value and size of the metal interconnect layer capacitors; separately extracting relationship data between voltage and capacitance value of the metal interconnect layer capacitors and between temperature and capacitance value of the metal interconnect layer capacitors by using a process device simulation tool, and add the relationship data to the relationship formula; and establishing a simulation model in accordance with the relationship formula of capacitance value, size, voltage and temperature. The method has improved modeling speed and reduced circuit design cycle. The model thereof can be applied to the analysis of the small size capacitors with reliability.

The invention provides a common assessment platform for the man-machine working efficiency of a control cabin of a helicopter and belongs to the technical field of man-machine working efficiency design of the helicopter. The device comprises a control cabin floor, and the control cabin floor is provided with seats of the control cabin of the helicopter, a central controlling table, an instrument board and a top controlling table; the instrument board, the central controlling table, the top controlling table and the seats are installed on the control cabin floor through supporting parts, and the height of the supporting parts can be adjusted; the two seats are symmetrically distributed along the axis of the control cabin, the bottoms of the seats are installed on the control cabin floor through horizontal slide rails, and seat faces are connected with backrests through rotation mechanisms. The common assessment platform can be applied to assessment of the view field, reachability and comfort of control cabins of different helicopters, according to the development demand of the control cabin of the helicopter and an assessment result of an actual pilot, input is provided for type design of the helicopter, the iteration time is shortened, and the assessment efficiency is improved.

The invention provides a wetland effluent ammonia nitrogen concentration prediction method and system based on a hybrid neural network, and the method comprises the steps: carrying out the principal component analysis and moving average processing of the obtained water quality environment data of the water inlet and outlet section of a constructed wetland, and obtaining input data; the number of indexes of the water quality environment data is not less than 8; constructing a training set and a test set by adopting input data, and constructing a prediction model by adopting a BP neural network; optimizing the weight and the threshold value of the BP neural network by adopting a genetic algorithm to obtain an optimized weight value and an optimized threshold value, and substituting the optimized weight value and the optimized threshold value into the neural network to complete the optimization of the neural network to obtain an optimized model; training the optimized model by adopting an LM algorithm to obtain an optimal water quality parameter prediction model; and predicting the ammonia nitrogen concentration of the constructed wetland to be predicted by using the test set. By adopting the technical scheme of the invention, the robustness and generalization ability of the model are improved, and the accuracy of effluent ammonia nitrogen concentration prediction is improved.

The invention discloses a moving target tracking and positioning method based on a single camera, and the method comprises the following steps: S1, carrying out the extraction of a moving target, carrying out the image collection of the moving target through the single camera, and the color space of a collected image being an RGB space; S2, identifying a moving object, specifically realizing the identification of the moving object through a digital image processing method; and S3, tracking and positioning the moving target, detecting the edge of the moving target through a Camshift algorithm and a neighborhood linear search method, and determining the relationship with the surrounding environment so as to carry out positioning processing. Compared with a traditional target tracking and positioning method, the iteration time of the tracking operation is reduced in the tracking time, the phenomena of frame loss and sudden change of the position of the moving target between two adjacent frames of collected images caused by the precision of the camera are effectively solved, and the method has the advantages of accurate tracking and positioning, high applicability, time saving and simple and convenient process.

The invention discloses a pothole detection method based on improved YOLOv3, comprising: S1, collecting pothole images through a visual acquisition system, obtaining a pothole data set after preprocessing, and the pothole data set includes the preprocessed pothole images ; S2, constructing an improved YOLOv3 pothole detection network model; S3, inputting the training data set of the pothole data set into the improved YOLOv3 pothole detection network model training, when the improved loss function approaches zero , obtain the parameter optimal solution of the improved YOLOv3 pothole detection network model; S4, input the pothole data set into the improved YOLOv3 pothole detection network model of the parameter optimal solution, and obtain the pothole Test results. The invention solves the problem that the detection of potholes must not only ensure real-time performance, but also further improve the accuracy rate.

The invention discloses an optical flow field calculation method. According to the method, current new iteration values of light streams in a lower neighborhood and a left neighborhood of a pixel point are already obtained when a light stream mean value of the pixel point is calculated through a four-neighborhood template, so that current iteration light streams of the pixel point can be obtained with the new value according to the calculated light stream mean value during calculation of the light stream mean value of the point. Through adoption of the method, errors are small, and the convergence speed of an algorithm is increased. The method is particularly suitable for situations with a large quantity of boundaries and large boundary gray scale gradients; the iteration times can be reduced greatly; and the iteration time is shortened. Compared with a conventional differential method, the method has the advantages of very high robustness, great shortening of the optical flow field calculation time, and high efficiency.

The invention discloses a method and system for graphically displaying a clock structure, relates to the technical field of chip clock design, and solves the technical problem that a text file containing clock information and a clock structure diagram cannot be freely and quickly converted. The main points of the technical solution are Read the clock information of all clock nodes under the root clock node through a scripting language, save the clock information into the first hash variable; create an html-based element for each clock, and write the clock information into the In the tag of the element; add a CSS style and a dynamic response event to the element, and generate a multi-level list diagram of the root clock node according to the first hash variable, and generate an html file. Fast and free conversion between text files containing clock information and clock multi-level list diagrams is realized, which greatly reduces the iteration time when analyzing and checking clock structures and modifying clock definition information when designing complex clock structure diagrams, and speeds up Clock Design Process.