56results about How to "Get the most out of performance" patented technology

The invention discloses a method for accelerating convolution neutral network hardware and an AXI bus IP core thereof. The method comprises the first step of performing operation and converting a convolution layer into matrix multiplication of a matrix A with m lines and K columns and a matrix B with K lines and n columns; the second step of dividing the matrix result into matrix subblocks with m lines and n columns; the third step of starting a matrix multiplier to prefetch the operation number of the matrix subblocks; and the fourth step of causing the matrix multiplier to execute the calculation of the matrix subblocks and writing the result back to a main memory. The IP core comprises an AXI bus interface module, a prefetching unit, a flow mapper and a matrix multiplier. The matrix multiplier comprises a chain type DMA and a processing unit array, the processing unit array is composed of a plurality of processing units through chain structure arrangement, and the processing unit of a chain head is connected with the chain type DMA. The method can support various convolution neutral network structures and has the advantages of high calculation efficiency and performance, less requirements for on-chip storage resources and off-chip storage bandwidth, small in communication overhead, convenience in unit component upgrading and improvement and good universality.

The invention discloses a hot-blast furnace pipe packing material and a preparation method therefor. The packing material is prepared from the ingredients in parts by weight: 100 parts of dry mattersand 10-25 parts of silica sol, wherein the 100 parts of dry matters comprise high bauxite, flint clay or mullite particles, silicon carbide particles, high bauxite or corundum powder, metal-nonmetal composite powder, activated alumina micropowder, calcined alumina micropowder, clay micropowder, carboxymethyl cellulose or hydroxymethyl cellulose and a curing agent. The packing material disclosed bythe invention is pollution-free to environments, is good in fluidity in construction time, is not prone to settlement, can be thoroughly packed into small slits during pressed-in repair, can be constructed at high temperatures and can also be cured at normal temperatures; the packing material can be tightly bonded to a material of a pressed-in area; the packing material has good strength propertyat all temperatures; the packing material has low linear shrinkage in service temperature intervals and has good thermal shock stability; and industrial wastes of other industries are utilized, so that the utilization ratio of resources is increased.

An embodiment of the invention discloses an image quality adjusting method and terminal. The method includes the steps: receiving image information and calculating dynamic frame rate D within a presettime; calculating an image quality grade J needing adjustment by an image quality grade adjustment algorithm according to the dynamic frame rate D; comparing a current image quality grade C with theimage quality grade J needing adjustment, and adjusting the image quality grade according to comparison results. The dynamic frame rate D is image frames transmitted per second within the preset time.If image quality is adjusted for the first time, the current image quality grade C is a recommended image quality grade T, otherwise, the current image quality grade C is an image quality grade obtained by previous image quality adjustment. The terminal receives image information, calculates the dynamic frame rate D, automatically adjusts the image quality according to the dynamic frame rate D, equipment performances can be sufficiently used, complicated operation is avoided, and user experience is improved.

Disclosed are an information processing method, retrieving method, device, user terminal and server, the retrieving method comprising: a first user terminal receives the address information transmitted by a second user terminal for saving an encrypted file of the user of the second user terminal; acquiring the private information, the non-private information and a shared key corresponding to the address information from a cloud server and a trusted server; acquiring the address information of a file to be accessed by retrieving the private information and the non-private information; acquiring from the cloud server a first encrypted file corresponding to the address information of the file to be accessed; decrypting, via the shared key, the first encrypted file to acquire the file to be accessed. The present invention introduces a trusted server during the information retrieval process, such that an encrypted file uploaded to a cloud server by a user terminal can be retrieved by a trusted third-party user terminal, thus taking full advantage of the cloud server performance while ensuring user file security.

The invention belongs to the field of composite material structure optimization methods, and discloses a fiber reinforced composite material structure optimization method based on Shepard interpolation. The method comprises the following steps of uniformly defining a series of discrete design points in a composite structure design domain, and building a continuous overall function to express a whole design domain fiber angle by using fiber angle values at the design points in a Shepard interpolation format; defining a fiber reinforced composite material structure optimal design problem, wherein a designing target is to minimize flexibility of the structure, and design constraints comprise a balance equation and upper and lower limits of [theta]i; dividing finite element grids to perform finite element analysis, and solving a displacement field; deducing sensitiveness of the flexibility of a target function with respect to a design variable, and computing the sensitiveness by using a finite element result; and updating the design variable by using an optimal algorithm until the target function is converged. According to the method provided by the invention, an optimization result fully utilizes material performance, manufacturing is facilitated, the design variable is reduced and optimization efficiency is relatively high.

The embodiment of the invention provides an automatic driving control system which comprises a sensor set, an executor set, a first controller and a second controller. The first controller is used forreceiving first sensing data sent by a first sensor, generating a first control command according to the first sensing data, sending the first control command to a first executor and executing presetfirst safety operation when the second controller breaks down. The second controller is used for receiving second sensing data sent by a second sensor, generating a second control command according to the second sensing data, sending the second control command to a second executor and executing preset second safety operation when the first controller breaks down. The safety of the automatic driving control system is guaranteed. The different controllers take charge of multiple automatic driving functions, and the performance of the two controllers can be fully used.

The invention discloses a multi-channel video recording and platform interaction method based on an Android system. The method comprises the following steps: carrying out image signal processing on multiple channels of original video data collected in real time, and converting the data into video data in a preset format and storing the obtained video data; calling a multimedia encoder / decoder of the Android system through an application program interface, and selectively carrying out video encoding / decoding processing on one or multiple channels of video data in the preset format to obtain corresponding high-definition video data; and storing the high-definition video data into a storage device and / or uploading the high-definition video data to a cloud server in a streaming media manner. The invention also provides a multi-channel video recording and platform interaction device based on the Android system. The method and system can realize compressed playback of multi-channel videos and does not influence application of other Android applications.

The invention discloses a camera shooting module and a control method for the camera shooting module. The camera shooting module comprises a first imaging assembly, a second imaging assembly and a sensor assembly. The first imaging assembly comprises a first diaphragm. The second imaging assembly includes a second diaphragm. The sensor assembly is located on the exit optical path of the first diaphragm and the sensor assembly is located on the exit optical path of the second diaphragm. When the camera shooting module is operated, the first diaphragm and the second diaphragm are alternately opened such that the sensor assembly alternately receives the first imaging beam generated by the first imaging assembly and the second imaging beam generated by the second imaging assembly. The technical scheme of the invention provides camera shooting module with simple and light structure and low cost.

The invention discloses a text file parallel uploading method and device. The method includes the following steps that: a text file to be uploaded is split into N data blocks, wherein N is an integer greater than 1; and N sub threads are started, and the N sub threads are utilized to simultaneously upload the N data blocks to a distributed file system according to a one-to-one corresponding relationship. According to the method of the invention, the text file to be uploaded is split into the N data blocks, and the N sub threads are utilized to simultaneously upload the N data blocks, wherein the N sub threads are one-to-one correspondence with working nodes, and therefore, compared with a method according to the which a whole text file to be uploaded is uploaded through one working node in the prior art, the method of the invention according to which the N working nodes are utilized to simultaneously upload the N data blocks, the performance of a whole cluster can be fully utilized, and high uploading speed can be achieved.

The invention discloses an active vibration control system of a flexible structure based on a voice coil motor and a method thereof, wherein the system comprises a main support rod fixed on the ground, a rail is fixed on the main support rod, left and right oblique rods on both sides of the rail are symmetrically fixed to the main support rod, left and right large displacement actuators adjacent to the left and right oblique rods are fixed on both sides of the main support rod when the rail is as a symmetry axis, both sides of the other end of the rail are symmetrically fixed with stationary fixtures on which fixed pulleys are mounted, and the top of the rail is provided with the flexible structure; the other ends of the left and right oblique rods are connected with positioning fixtures on which hole bolts are mounted, the fixed pulleys, the hole positions of the hole bolts and the left and right oblique rods are located on the same plane, and a rigid rope passes through the left large displacement actuator, the fixed pulleys and the holes of the hole bolts; besides, the invention further discloses a control method of the system; the active vibration control system outputs large displacement and greater control force through the voice coil motor, a control direction is shifted by the pulleys such that the active modal vibration control of the flexible structure can be realized.

The invention provides a peak clipping and valley filling method applied to a power supply of a communication base station in the field of power supply management. The peak clipping and valley fillingmethod comprises the steps of S1, sending a power failure command to a battery manager during power failure; S2, detecting whether a relay is closed or not, and if so, prompting a fault; if not, entering the step S3; S3, judging whether the electric quantity of the lithium battery is less than 15% or not, and if so, prompting a fault; if not, sending a relay closing command, and entering the stepS4; S4, judging whether the relay is closed or not, and if so, entering the step S5; if not, prompting a fault, and entering the step S3; S5, allowing the lithium battery to supply power to the direct-current load, judging whether power is on or not, and if not, entering the step S6; if yes, conducting peak clipping and valley filling according to the current time period; S6, judging whether theelectric quantity of the lithium battery is less than 15% or not, and if so, prompting a fault; and if not, entering the step S5. The method has the advantages that the performance of the lithium battery is fully utilized, and the reliability of the power supply of the communication base station is greatly improved.

The invention relates to the field of retailing operation modes, and specifically relates to a distributed background service system and a distributed background service method based on a vending machine. The system comprises a data acquisition unit, a central processing unit and a background application service unit, wherein the data acquisition unit is used for acquiring real-time work state information of the vending machine; the central processing unit is used for integrating and processing the acquired real-time work state information, and performing packaging and sending; the backgroundapplication service unit is used for performing service division for the integrated and processed work state information, and storing services according to service types. With the system and the method provided by the invention, load balance can be realized, and pressure on a background server can be reduced; by subdividing the services into atoms, an architecture is clear and convenient to maintain; moreover, when one host (service) crashes down, normal operation of the other services cannot be influenced. Moreover, machines with different configurations are selected for different services, thus, machine performances are used fully.

The invention belongs to the technical field of robots and high-grade numerical control machine tools, and particularly relates to an interval adaptive planning method for five-axis machining feed speed. Based on application of a look-ahead window, the influence of the out-of-window curve morphology on the current interval feed speed planning is cut off; allowable feed parameters of a current interval and a look-ahead interval are calculated based on a differential principle; and the feed speed profile of the current interval is planned according to the allowable feed parameters of the previous interval, the current interval and the look-ahead interval, so that regional look-ahead planning of the overall feed speed of the long spline curve path is achieved. The calculation efficiency is high, and the machine tool performance can be fully utilized on the premise that feed shaft driving capacity constraints are met.

The embodiment of the invention discloses a global optimal feeding speed planning method for a cradle type machine tool. According to the scheme, the method comprises the steps that the speed, the acceleration and the jerk are restrained in a machine tool coordinate system, and the bow height error is restrained in a workpiece coordinate system. Then, the optimal feeding speed planning problem is converted into an optimal control problem, and it is indicated that the optimal solution of the optimal control problem is 'bang-bang' control. And finally, an optimal solution of the optimal control problem is solved by adopting an iterative control vector parameter method, and the solution meets 'bang-bang' control. According to the method, various constraints including motion constraints and geometric error constraints in the whole machining process of the cradle type machine tool can be fully considered, and the whole cutting task is completed in the shortest time, so that the performance of the machine tool can be fully utilized.

The invention discloses an integrator of a variable integral time constant. The integrator comprises an operational amplifier and is characterized in that: a voltage input end is connected to an inverting input end of the operational amplifier after passing through an option switch and an option circuit comprising a plurality of resistances with different resistance values; a non-inverting input end of the operational amplifier is grounded; the option switch and the option circuit comprising a plurality of resistances with different resistance values are connected between the inverting input end and an output end of the operational amplifier. The integrator of the variable integral time constant, which is provided by the invention, can switch over an integrating resistance or an integrating capacitance according to the integrator output condition last time (or in a period of time) so that the output amplitude of the integrator is not too small and Signal-to-Noise can be improved, or the integrator prevents errors caused by output saturation and makes full use of performance of the integrator.

The invention discloses a steel corbel pre-embedded structure and a construction method thereof. The steel corbel pre-embedded structure comprises a concrete column, a corbel transverse steel plate, acorbel transverse connection steel plate and two corbel vertical steel plates. The two corbel vertical steel plates are inverted L-shaped plate structures arranged symmetrically, the vertical corbeltransverse connection steel plate is connected between the middle parts of the corbel vertical steel plates in a welded mode, one ends of the corbel vertical steel plates extend into the concrete column, and the other ends of the corbel vertical steel plates are located on the side edge of the concrete column. The corbel transverse steel plate is connected to the top ends of the outer sides of thecorbel vertical steel plates in a welding mode, pre-embedded steel pipes are arranged in the concrete column, split finish rolling threaded steel is arranged in the pre-embedded steel pipes, the thread steel is fixed through screw nuts after penetrating through preformed holes of the corbel transverse connection steel plate. Through the pre-embedded steel pipes and the split finish rolling threaded steel, and the tensile stability of the structure is ensured; and through a pre-embedded steel plate box, the corbel steel plates are inserted into the pre-embedded steel plate box, the steel plates is used for bearing the shear force to ensure that properties of materials are well used, and good using effects are obtained.

The invention discloses a self-reset viscosity composite damper which comprises a viscosity damper, a guide rod, a U-shaped shape memory alloy element and a connecting frame. A cylinder body of the viscosity damper is fixedly connected with the connecting frame. One end of the guide rod penetrates the connecting frame and is connected with a piston rod of the viscosity damper. The branch end of the U-shaped shape memory alloy element is connected with the guide rod and the connecting frame. The shape memory alloy and the viscosity damper are used for forming the composite damper, characteristics of the self-reset damper can be developed in the initial stage of an earthquake, the characteristics of the viscosity damper are developed when the earthquake is strong, enough energy for consumption is provided, the characteristics of the self-reset damper is also developed before the earthquake is over, almost no residual deformation exists on the structure obtained after the earthquake, theearthquake resisting performance of the structure and the function recovery capability of the structure after the earthquake can be overall improved, and the self-reset viscosity composite damper hasa wide application prospect in the field of building structures and bridge engineering.

The invention discloses a ribbed reinforced concrete prefabricated board, comprising a reinforced concrete base plate, a longitudinal rib and prestressed reinforcing steel bars, wherein through holes are drilled at two sides of the longitudinal rib of the reinforced concrete prefabricated board, and grooves are formed at the positions, corresponding to the through holes, of the base plate. A floor slab is formed by assembling a plurality of the reinforced concrete prefabricated boards, the hole-penetrating reinforcing steel bars penetrate through the holes at two sides at the lower part of the longitudinal ribs of the prefabricated boards and are placed on the grooves on the base plates, meanwhile, a plurality of reinforced concrete prefabricated boards are connected, and concrete is poured then for flatness. Because the placement positions of the hole-penetrating reinforcing steel bars are lower, the performance of the hole-penetrating reinforcing steel bars can be used fully, and the grooves on the base plates can be used for enhancing the cementation between the prefabricated boards and the concrete, so that the integral performance of the floor slag is enhanced.

The invention relates to the technical field of integrated circuit computer aided design, and concretely relates to a parallel processing method and system for single fine granularity in capacitance extraction of random walk. The method comprises the following steps: obtaining interconnection line information, and cooperating with a process file to obtain a physical three-dimensional structure of an integrated circuit; configuring a plurality of threads and the converging precision according to user requirements and the number of processor operation cores; constructing a closed and wrapped Gaussian surface for each conductor; calculating probability density according to the area of the Gaussian surface; selecting a plurality of points from the Gaussian surface according to the probability density; putting the selected points into an operation queue as a single operation unit; enabling each thread to sequentially extract points on the Gaussian surface from the operation queue to execute random walk, and if the random walk reaches the surface of the conductor or the boundary, completing one-time random walk, and storing a random walk result until the operation queue is empty; and if the random walk result reaches convergence precision, stopping capacitance extraction, and outputting an extraction result. The method is high in parallelism and high in expansibility.

The invention discloses an index migration method and device and electronic device. The method comprises the steps: obtaining a first request according to a target task queue, wherein the first request is used for updating a first index of target data, and the target data is incremental data received by a search engine in the index migration processing process; according to the first request, updating the first index, and according to execution states of all requests in the first request, obtaining first state information; under the condition that the first state information shows that the first request is successfully executed, executing second index updating processing of the target data, wherein the structure of the second index is different from that of the first index. According to the method, smooth migration of the data on the new and old indexes can be conveniently and efficiently realized.

The invention discloses a method and a system for deploying a Web server. The method comprises the following steps: deploying Nginx proxy services on a main server and N business servers respectively;generating an independent virtual IP on the main server and the N service servers by utilizing a Keep Alive mechanism; and establishing a mapping relationship among the virtual IP, the main server and the N service servers. The system comprises a main server, N service servers, a deployment module, a virtual IP generation module and a mapping relationship establishment module. According to the method and the system, the stability of the Web server can be greatly improved, and load balancing and downtime hot standby can be better realized.

The invention discloses a bridgehead car jump restoration and precaution method based on load reduction. The problem that car jump at bridgeheads exists in the prior art is solved, and the problems that methods adopted at present has high requirements for materials, a long curing time is needed during forming, the construction period is long, and the influence on traffic operation is large are solved. The method comprises the steps that holes are drilled at intervals in the original road face of a bridgehead; hollow barrels are buried in the holes, and the holes are backfilled until the holes are flush with the original road face of the bridgehead; and a road face is additionally laid on the original road face of the bridgehead to restore the bridgehead line shape. The method has the advantages that through the load reduction method, the strength and performance of the original road face are fully used to achieve the purpose of load reduction, and then later settlement is effectively controlled and treated; the hollow barrels are buried in the holes, the curing time needed by slurry or foam concrete is omitted, the displacement work amount is reduced, and the influence period on the traffic is short; and machines are not needed for mounting, and construction is simpler, more convenient and more economical than that of foam concrete.

The invention discloses a rapid slurry dehydration device and a dehydration method thereof. The device comprises a hollow outer cylinder; a plurality of water seepage holes are uniformly formed in the side wall of the hollow outer cylinder; geotechnical cloth is attached to the inner side wall of the hollow outer cylinder; a bottom plate is closed and fixed at an opening in one end of the hollow outer cylinder; one end of a cross plate is fixedly connected with the bottom plate in an inserted way; a partition plate group comprises at least one set of partition plates; the number of the isolation plates in one set is four, each isolation plate is a long-strip-shaped rectangular plate, one end of each isolation plate is connected with the bottom plate in an inserted mode, the two side edges of each isolation plate are attached to the two adjacent side walls of the cross plate, and a three-dimensional structure with the rectangular cross section is defined by the four isolation plates; and a high-performance water-absorbent resin bag is packed in a gap between the inner wall of the hollow outer cylinder and the outer side wall of the outermost set of isolation plates. The invention provides a rapid dehydration device for high-water-content slurry and a dehydration method of the rapid dehydration device, aiming at the problems of low efficiency, high energy consumption, easiness in abrasion of equipment and incapability of achieving economization and sustainability of a slurry dehydration method at the present stage.

The invention provides a concrete admixture prepared from sulfur-fixing ash and sulfur-fixing residue, wherein the preparation method comprises: pretreatment, alkali excitation, wall breaking shaping(grinding), sorting, and homogenization, and the prepared concrete admixture comprises, by weight, 50-70% of sulfur-fixing ash, 30-50% of sulfur-fixing residue, and 1.0-1.5% of an alkaline activator.According to the present invention, the concrete admixture prepared from sulfur-fixing ash and sulfur-fixing residue is provided, such that the large-doping and large-batch use of sulfur-fixing ash and sulfur-fixing residue in concrete is achieved, the waste is turned into treasure, the solid waste is subjected to large-scale and resource utilization, and the unification of production benefits, environmental benefits, economic benefits and social benefits is finally achieved.

The invention discloses a power amplifier monitoring method and a power amplification circuit, and the method comprises the steps: collecting an analog voltage group and an analog current group of a power amplifier in real time through a collection module, and transmitting the analog voltage group and the analog current group to a conversion module; performing voltage analog-to-digital conversion on the analog voltage group through a conversion module to obtain a digital voltage group, performing current analog-to-digital conversion on the analog current group to obtain a digital current group, and transmitting the digital voltage group and the digital current group to a monitoring module; the monitoring module outputs the working parameters of the digital voltage group and the digital current group to obtain the parameter group and judges the threshold values of the working parameters in the parameter group to obtain the monitoring result, and the monitoring result shows whether the power amplifier works in the safe area or not. The working state of the power amplifier is accurately monitored, and the performance of the power amplifier is fully utilized.

The invention provides a heterogeneous computing architecture camera system and an image processing method, the system comprises an image acquisition device and an image processing host, the image acquisition device comprises an image sensor and a field programmable gate array module, and the field programmable gate array module receives data output by the image sensor for preprocessing; the imageprocessing host comprises a central processing unit and an image processor, and the image processing host receives the preprocessed data output by the field programmable gate array module. The methodcomprises the steps that an image sensor of an image acquisition device acquires data and outputs the data to a field programmable gate array module, and the field programmable gate array module receives the data output by the image sensor for preprocessing including color correction, image enhancement and data compression; and the image processing host receives the preprocessed data output by the field programmable gate array module and processes the preprocessed data again. The system complexity can be simplified, and the image quality and the processing speed are improved.

The invention discloses a process management method and a process management device, and relates to the technical field of process management. The number of each process type of process instantiation corresponding to a logic node is obtained, and when the CPU occupancy rate of business process running on one CPU core in the logic node is greater than a first preset occupancy rate, a process of which the type is the same as that of the business process is loaded on other CPU cores according to the number of process instantiation of the business process. Due to the fact that the process with the same type is loaded into the other CPU cores, the performance of CPU is sufficiently utilized, and the method is convenient to realize.

An embodiment of the invention provides a gatekeeper load balancing method and device based on file transmission. The gatekeeper load balancing method and device based on file transmission are applied to a gatekeeper cluster. The gatekeeper cluster comprises at least two gatekeepers, wherein each gatekeeper is identified as a node, one gatekeeper is identified as a control node, and each remaining gatekeeper is identified as a data node. The method comprises the following steps of: scanning an external network side server through a control node, and determining target file data to be transmitted; acquiring target file data needing to be transmitted by each node according to the target file data to be transmitted, and sending the target file data to each node; the control node and the data node pull a target file from an external network side server according to the target file data needing to be transmitted by each node; and the control node and the data node upload the pulled target file to an internal network side server. The control node can control each data node to pull and transmit the same number of target files, so that the load balance of the gatekeeper cluster is realized, and the overall transmission capability of the gatekeeper is improved.

The invention discloses a preparation method of a lead graphene battery negative electrode plate. The preparation method is characterized by comprising the following steps: S1: carrying out graphene oxide modification, heating polar nitrogen containing organic solvent to 50-80 DEGC, adding graphene oxide, continuously stirring for 3-5h, and separating and drying to obtain modified graphene oxide; S2: dispersing the modified graphene oxide into a metal salt solution; S3: transferring the solution obtained in the S2 into a hydrothermal kettle, carrying out hydrothermal reaction for 3-8 hours at the temperature of 150-170DEG C, and obtaining the composite material of reduced graphene metal oxide; S4: dispersing the composite material into a sulfuric acid solution to obtain graphene sulfuric acid dispersion liquid; S5: mixing lead powder, barium sulfate, lignin, fibers and the graphene sulfuric acid dispersion liquid into paste, drying and solidifying to obtain the lead graphene battery negative electrode plate. Through the S1, a graphene surface is subjected to functionalization in advance, then, the graphene and the metal oxide are compounded, and the obtained composite material can fully perform the function in batteries.