129 results about "Vector mode" patented technology

In addition to the usual modes of SIMD processor operation, where corresponding elements of two source vector registers are used as input pairs to be operated upon by the execution unit, or where one element of a source vector register is broadcast for use across the elements of another source vector register, the new system provides several other modes of operation for the elements of one or two source vector registers. Improving upon the time-costly moving of elements for an operation such as DCT, the present invention defines a more general set of modes of vector operations. In one embodiment, these new modes of operation use a third vector register to define how each element of one or both source vector registers are mapped, in order to pair these mapped elements as inputs to a vector execution unit. Furthermore, the decision to write an individual vector element result to a destination vector register, for each individual element produced by the vector execution unit, may be selectively disabled, enabled, or made to depend upon a selectable condition flag or a mask bit.

A speech recognition system for transforming an acoustic signal into a stream of phonetic estimates includes a frequency analyzer for generating a short-time frequency representation of the acoustic signal. A novelty processor separates background components of the representation from region of interest components of the representation. The output of the novelty processor includes the region of interest components of the representation according to the novelty parameters. An attention processor produces a gating signal as a function of the novelty output according to attention parameters. A coincidence processor produces information regarding co-occurrences between samples of the novelty output over time and frequency. The coincidence processor selectively gates the coincidence output as a function of the gating signal according to one or more coincidence parameters. A vector pattern recognizer and a probability processor receives the gated coincidence output and produces a phonetic estimate stream representative of acoustic signal.

The invention discloses an automatic question answering method, an automatic question answering system and a method for constructing a question answering case base. The automatic question answering method mainly includes vectorizing questions submitted by users by using the question answering case base which stores question answering cases in a vector mode to generate question vectors, searching answer knowledge point IDs with satisfactory similarities by using the question vectors and the question answering case base containing case vectors, and obtaining answer content output to users from a knowledge point document case by using the answer knowledge point IDs. According to the system and the methods, the cost can be reduced, and the working efficiency can be improved.

The invention relates to an abnormal cardioelectric recognize method based on ultra-complete character. Wherein. it first uses secondary sample wavelet to process R-point check on the cardioelectric data, segments and pretreats the cardioelectric data based on the R-point position; then uses independent component method and disperse wavelet convert method to extract one ultra-complete character group from each heart beat, and contract the character via communication method; at last uses the vector method to train the extracted character to obtain one vector mode, to automatically recognize and classify the new heart beat data.

The system with a PC software part and FPGA hardware comprises: a bottom layer to transfer information on physical channel, a middle layer to pack and unpack data frame, and a top layer for signal input / output. Wherein, it sends ahead the expected result into the hardware for storage. This invention improves simulation speed greatly with well accuracy.

An improved superscalar processor. The processor includes multiple lanes, allowing multiple instructions in a bundle to be executed in parallel. In vector mode, the parallel lanes may be used to execute multiple instances of a bundle, representing multiple iterations of the bundle in a vector run. Scheduling logic determines whether, for each bundle, multiple instances can be executed in parallel. If multiple instances can be executed in parallel, coupling circuitry couples an instance of the bundle from one lane into one or more other lanes. In each lane, register addresses are renamed to ensure proper execution of the bundles in the vector run. Additionally, the processor may include a register bank separate from the architectural register file. Renaming logic can generate addresses to this separate register bank that are longer than used to address architectural registers, allowing longer vectors and more efficient processor operation.

The invention relates to an optical flow characteristic vector mode value and angle combination micro expression detection method based on an interested area. According to the method, firstly, a micro expression video is pre-processed to acquire a micro expression sequence, key face characteristic points are further extracted, and a face interested area having the best effect is found out according to motion characteristics of different expression FACS motion units; optical flow characteristics of the interested area are extracted. The optical flow vector angle information is firstly introduced, an optical flow vector mode value and the angle information are acquired through calculation, through combination of the two, a more comprehensive and more judicious characteristic detection micro expression segment can be acquired; the optical flow mode value and the angle information are combined, the threshold is determined according to magnitude of the optical flow mode value, and a number shape combination method is utilized to vividly and visually acquire the micro expression segment. The method is advantaged in that micro expression detection efficiency is substantially improved, optical flow characteristic vector extraction is carried out only in the key face area, computational complexity is reduced, the consumption time is shortened, and high robustness is realized.

A processor that can execute instructions in either scalar mode or vector mode. In scalar mode, instructions are executed once per fetch. In vector mode, instructions are executed multiple times per fetch. In vector mode, the processor recognizes scalar variables and vector variables. Scalar variables may be assigned a fixed memory location. Vector variables use different physical locations at different iterations of the same instruction. The processor includes circuitry to automatically index addresses of vector variables for each iteration of the same instruction. This circuitry partitions a register into a vector region and a scalar region. Accesses to the vector region are automatically indexed based on the number of iterations of the instruction that have been performed.

The invention discloses an information pushing method and device, terminal equipment and a storage medium. The method comprises the steps that historical behavior information of a user is obtained; the historical behavior information is analyzed and filtered, and user keywords are obtained; the user keywords are trained in a word vector mode, and user tags are determined; on the basis of a K-Meansaggregation algorithm, the user tags are classified, and an original user tag group is obtained; the original user tag group is sorted according to the historical behaviour information historical browsing history of the user, and a target user tag group is obtained; on the basis of the target user tag group, preference information of the user is obtained; service tags corresponding to the preference information are obtained from a preset service tag library, and service information corresponding to the service tags is pushed to the user. According to the technical scheme, the intelligent level of service information pushing and the popularization efficiency of the service information are effectively improved.

The invention discloses a word vector model incremental study method. The super parameter used in the method comprises: a vector dimension, a number scope negative data, and a text window length; the word vector mode is updated dynamically to complete the vector model optimization and therefore complete the incremental study through initializing new words that appear in the additional text and sampling counterexample based on the historical glossary word-list; the technical scheme study can avoid to study repeatedly, historical data the calculating complexion degree is greatly reduced; the scheme can maintain the higher study efficiency so as to satisfy the efficiency requirement of the online system.

The invention discloses a method for realizing general demodulation of different modulating signals. The method comprises the following steps: firstly, using the direct numerical frequency synthesis method to generate two routes of local orthogonal carrier signals by expanding the rotation mode under the circumferential coordinate in the coordinated rotation digital computer (CORDIC) algorithm ofa convergence domain to synthesize equidirectional subcircuit data xI (t) and orthogonal subcircuit data xQ (t) with receiver terminal output signals; then filtering secondary harmonic components by a low pass filter; afterwards, inputting two routes of signals into a general demodulation module of the vector mode under the circumferential coordinate in the CORDIC algorithm of the convergence domain; finally, selecting a square root operation value or an arc tangent operation value according to different modulation modes to obtain demodulation information by filtering or differencing. In the invention, various modulation modes are integrated in the same hardware, so that the method has the characteristics of good flexibility and outstanding versatile performance; and meanwhile, the complex arithmetic operation is transformed into the displacement and add operation, as a result, the method can greatly reduce the complexity of hardware realization and has good carrier frequency mismatching resistance property.

The invention relates to a method and system for judging spatial relationship based on gridded GIS. Step 1, input the graphics represented by the geometry represented by the vector mode, cover the graphics by the grid unit, the grid unit takes the element ID that covers the graphics as the element index, and the grid unit takes the coordinate value Position index, which divides the geometric expression forming the index into blocks, and each grid unit has the information of the actual occupied area of ​​the graphics covered by it in the grid unit; step 2, two compressed grid units In the covered graphics, the grid units with the same coordinates correspond to each other, and the information of the actual occupied area of ​​the corresponding two grid units is compared, and the two grid units covered by the grid unit are obtained in a vector form. The result of judging the GIS spatial relationship expressed by the represented geometry.

The invention provides an initial positioning method for a permanent-magnet synchronous motor. The method includes that a rotary transformer, a rotation-digit conversion module and a level conversion module are additionally arranged, wherein a stator winding and a rotor winding of the rotary transformer are respectively fixed on a stator and a rotor of the permanent-magnet synchronous motor, the stator winding comprises a primary winding and a secondary winding, the rotation-digit conversion module provides exciter signals for the primary winding and converts envelop signals output by the secondary winding into absolute angle signals, the rotation-digit conversion module communicates with a main processor of a system through the level conversion module, when the system is powered on initially, the rotor stops in the position of an angle of 0 degree, and the main processor acquires the initial angle A0 of the motor, then the permanent-magnet synchronous motor runs in a closed-loop vector mode, the main processor acquires an absolute angle Ax of the motor, and the position angle of the rotor of the permanent-magnet synchronous motor satisfies the equation that theta=Ax-A0, every time the system is powered off, the current position angle of the rotor is stored, and when the system is powered on again, the current position angle of the motor is taken as an initial angle.

The invention provides an autonomous underwater vehicle controlled by double propellers in a vector mode and provided with a foldable antenna. The autonomous underwater vehicle comprises an upper fairing and a lower fairing, the upper air guide sleeve and the lower air guide sleeve are connected to form a cabin body, a main cabin and power supply cabins arranged on the two sides of the main cabinare installed in the cabin body, a plurality of camera structures are arranged at the head of the main cabin, the tail part of the main cabin is connected with a tail part propelling mechanism, an antenna structure is arranged at the tail end of the main cabin, the tail propelling mechanism comprises propellers symmetrically arranged on the two sides of the tail of the cabin body, the two propellers are both connected with a first underwater steering engine which drives the two propellers to synchronously rotate on the vertical space plane and then adjust the pitching state, steering is controlled through the difference rotating speed of the two propellers, and the first underwater steering engine is installed in the cabin body and connected to a rear end cover of the main cabin.

A synchronization pattern generating unit generates a synchronization pattern required for a clock recovery unit which has been built into a DUT to maintain a link with an external circuit. A gate signal generating unit generates a gate signal which is asserted in a period in which a vector pattern is to be supplied to the DUT. In a first mode, a pattern selecting unit is configured such that it outputs the vector pattern during a period in which the gate signal is asserted and outputs a fixed output level during a period in which the gate signal is negated. In a second mode, the pattern selecting unit is configured such that it outputs the vector pattern during a period in which the gate signal is asserted and outputs the synchronization pattern during a period in which the gate signal is negated.

The invention discloses a control method of a modular multilevel converter based on a power frequency fixed switching frequency. The control method includes the steps that the modulation strategy that impulse waveforms are moved periodically is adopted, so that the characteristics of four impulse waveform parameters in one group are determined; specific values of the impulse waveform parameters are determined by referring to the latest level modulation strategy, fine adjustment is conducted on the impulse waveform parameters by the utilization of an impulse waveform parameter vector mode and according to an active energy equipartition principle, an impulse waveform parameter table is compiled, and under the condition that the total distortion rate of voltage harmonic is guaranteed, and automatic balancing of module absorbing power is achieved. A modulation ratio is determined by detecting the actual value of a power grid voltage and the actual value of a power grid current, the table is looked up according to the modulation ratio to obtain the impulse waveform parameters, then the impulse waveforms are moved periodically according to the corresponding principle and the number of sub-module cascades to obtain sub-module switching signals when capacitor voltage control does not exist, the control over the modular multilevel converter based on the power frequency fixed switching frequency is realized, and good reference value is provided for engineering application.

The invention discloses a method for realizing software and hardware co-simulation platform based on network, belonging to the technical field of SOC simulation and verification, which comprises a PC terminal for running the software part, a target board for simulating the hardware part, and a debugger for connecting the PC terminal and the target board mutually. The invention has the advantages that the simulation support to Vector mode and Co-simulation mode can be realized on a unitive multi-mode simulation system; the occupation on I / O resources is reduced greatly due to the communication between the debugger and the target board adopting serial-parallel conversion method; and the software part is communicated with the hardware part via Ethernet to support remote co-simulation.

The invention discloses a vector mode data transmission system and method based on assisted localization technology of sensing components. With the assisted and matched localization technology of built-in position and movement sensing components on various hand-held electronic devices, an intuitive 2D or 3D man-machine control interface is produced, and a user can transmit a file from a source device to a target device with an intuitive gesture. The vector mode data transmission method based on the assisted localization technology of the sensing components comprises the following steps that: (1) initial communication environments of a plurality of devices are set, and a mediation server is determined; (2) the plurality of devices acquire position information of various devices through localization technology and synchronize in-time position information together with the mediation server; (3) after a source device among the plurality of devices acquires localization information of the various devices after localization conversion via the mediation server, a user uses the localization information of the various devices after localization conversion to specify a target device for the file transmission; and (4) the file is transmitted to the target device.

The invention discloses a traffic condition analyzing device based on a vector mode. A road section between two adjacent intersections is taken as a unit road section. A collecting module, a normalized module and a comparison module are arranged. The collecting module is used for acquiring five traffic condition parameters of the unit road section at current time bucket; the normalized module is used for performing short-time prediction and normalization on the traffic condition parameters respectively and constituting a five-dimensional feature vector; and the comparison module is used for comparing the five-dimensional feature vector with various preset reference value respectively and the basic traffic condition corresponding to the reference value with the biggest similarity is determined as the current traffic condition of the unit road section. The device and the method have a simple model mode and low computation complexity. Meantime, the possible traffic change trend is the key consideration, thus the traffic condition analysis can have actual and practical effects. The device and the method have the advantages of low installation and maintenance cost, high working efficiency, high precision, strong adaptability and the like. The invention also discloses a traffic condition analyzing method based on a vector mode.

The embodiment of the invention discloses a codebook sub-sampling method applied to an LTE-A system and a device thereof. According to the system, a precoding matrix W is obtained by multiplication of a matrix W1 and a matrix W2 as well as the relation between the W and W1 and W2 is expressed by the following formula: W= W1W2; codebooks of the W, the W1 and the W2 are respectively marked as C, C1 and C2; and r means rank indication. The method comprises: sub-sampling is carried out on the codebook C, so that the size of the codebook C after the sub-sampling is equal to or is less than 4 bits. During the sub-sampling process, code words that are distributed uniformly are extracted from the codebook C; and a portion of or all code words are in a DFT vector mode so as to be suitable for a uniform linear array; and the rest of the code words are suitable for a cross polarization linear array.

The invention relates to a wave water surface sky diffuse reflection light polarization field simulation method. The method has the following steps: a celestial coordinate system describing a sky light polarization field is established; sun point location and observation point location are designed in the celestial coordinate system; a sky light scattering angle is calculated in spherical geometry; degree of polarization of sky light at any point is calculated according to a semi-analytical Rayleigh scattering model; a polarization angle of sky light is calculated according to the sun point location and sky observation point coordinates via a vector mode; the polarized reflectance and the degree of polarization of wave water surface sky diffused light are calculated; and the polarization angle of water surface reflected sky light is obtained according to information of sky light degree of polarization in combination with an STOKES vector and a trigonometric function relationship so that calculation of the polarization state of water surface reflected sky light is realized. Sky light can be incident at any polarization state, and the polarization state of the reflected light of a static water surface and the wave water surface can be realized so that polarization characteristics of a natural water surface can be really described.

The invention discloses a full-automatic laser marking system, which comprises a controller used for controlling a galvanometer system, a laser marking mode connected with a peripheral equipment control interface according to a requirement to uniformly identify and acquire different file formats for input of a digital camera or a scanner, and a laser optical system capable of identifying and acquiring the input of the digital camera or the scanner. By adopting a mode of processing graphics by an advanced graphics system in a computer, the full-automatic laser marking system has high drawing efficiency, high graphic accuracy, and the function of freely amplifying and reducing the graphics without distortion or deformation, the marking quality and marking speed are greatly improved by a vector mode, and the full-automatic laser marking system has the function of automatically correcting image distortion, can mark precise images, has a wide application range, can perform vector marking and dot matrix marking, ensures adjustable marking range, and has the advantages of high response speed and high marking speed.

A processor that can execute instructions in either scalar mode or vector mode. In scalar mode, instructions are executed once per fetch. In vector mode, instructions are executed multiple times per fetch. In vector mode, the processor recognizes scalar variables and vector variables. Scalar variables may be assigned a fixed memory location. Vector variables use different physical locations at different iterations of the same instruction. The processor includes circuitry to automatically index addresses of vector variables for each iteration of the same instruction. This circuitry partitions a register into a vector region and a scalar region. Accesses to the vector region are automatically indexed based on the number of iterations of the instruction that have been performed.

The invention discloses a resistance-changing-type single-fault-source diagnosing method for a mine ventilation system based on air volume characteristics. A training sample of the fault and air volume relation is generated through a mine ventilation simulating system (MVSS), a classification model and a regression model based on a support vector machine (SVM) are built, and the fault position andequivalent air resistance are diagnosed. The topology structure of a ventilation system network is complex, change of the air resistance of one of branches of the ventilation system network gives rise to change of the air volume of the branch and the ventilation network of a partial area, the air volume serves as a vector mode so as to distinguish the topology position where a resistance-changing-type fault occurs, and the single-fault-source SVM diagnosing method is put forward to diagnose the resistance-changing-type single fault source position and fault degree of the mine ventilation system; a final decision model is built according to the method, the method is simple, human interference is less, connection between data is objectively reflected, and the method is suitable for solvingthe essential non-linear problem; and the model is determined by a support vector, data dimension disaster is avoided, and good robustness is achieved.

A system for the efficient use of CORDIC rotators is provided. The system includes at least one first rotator operating in vector mode and includes a sign storage buffer. The system also includes at least one bit inverter coupled between the first rotator and a second rotator, the second rotator operating in rotation mode.

A system for baseband amplitude limiting is provided. The system includes a CORDIC vector mode rotator device having an angle accumulator in addition to a gain device and limiter. The system also includes a second CORDIC rotator operating in rotation mode, and is coupled to the limiter.

The invention relates to a face recognition method based on a manifold self-adaptive kernel. The method comprises the specific steps that a face image is represented in a vector mode through a representation method based on the appearance; the graph Laplacian is calculated; a non-parameter kernel matrix K is calculated; the graph Laplacian and the non-parameter kernel matrix K are calculated so that a manifold self-adaptive kernel function tightly related to the face data can be obtained; an optimized objective function of the kernel NMF is constructed, a Lagrange function is built, and a multiplicative updating rule is obtained; the kernel matrix calculation process is achieved through the low rank approximate technology of a Nystrom method; the optimized W and the optimized V are calculated, and for a new tested face image Ztest, low dimensional characteristic representation after the dimension reduction of the kernel NMF is obtained; an optimized objective function of an SVM is built; an optimized solution alpha is calculated. The face recognition accuracy of the manifold self-adaptive kernel is better than that of the existing algorithm, dimension reduction serves as the preprocessing algorithm, and then the performance of the face recognition algorithm is enhanced effectively.

The invention discloses a method and system for processing unstructured data. The method for processing the unstructured data comprises the following steps that multiple feature templates comprising keywords are arranged; a database where multiple columns of unstructured data are stored is scanned through each feature template, whether the content which is identical to each feature template and is recorded in the corresponding column of unstructured data exists in the unstructured data or not is judged, and each feature template with a positive judgment result is recorded as the feature template matched with the corresponding column of unstructured data; multiple template vectors corresponding to the multiple columns of unstructured data one to one are generated and each template vector is provided with multiple dimensions corresponding to the feature templates one to one. According to the method and system for processing the unstructured data, the unstructured data are processed through the feature templates so that the unstructured data can be in a vector mode, the following calculation processing is conducted on the template vectors, calculation resources needed for data analysis of the unstructured data are reduced, and the time needed for data analysis of the unstructured data is shortened.

The invention provides a deep text matching method and device based on unregistered word processing. When the depth matching model is trained, word segmentation is carried out on input sentences in aword vector auxiliary word vector mode. Words obtained through word segmentation of all the sentences are continuously subjected to word segmentation processing and expressed through word vectors, andthen a result of connection of the word vectors and the word vectors serves as the sentence quantity of the input sentences. Alternatively, the input statements are represented by initial word vectors, similarity weighting processing is performed on the initial word vectors, weighted word vectors are generated. Word segmentation processing is performed on the statements. The statements are represented by word vectors, and then the result of connection between the word vectors and the word vectors is used as the statement vector representation of the input statements. Due to the fact that newcharacters which do not appear in the training set rarely occur, interference of unlogged-in words on the model matching effect can be avoided to a certain extent in the mode that the characters and the word vectors are combined, and text matching precision of the model can be improved.