108 results about "Beamforming algorithm" patented technology

The present invention relates to a beamforming device in an orthogonal frequency division multiplexing access system, and a method thereof. In the method, a switched beamforming algorithm is applied to perform downlink beamforming, and an adaptive beamforming algorithm is applied to perform uplink receiving beamforming. In addition, a half wavelength array antenna is used to simultaneously allocate the same resource to two terminals separately positioned in two different spaces, so as to support spatial division multiple access (SDMA). In this case, since a base station uses channel status information at transmitter (CSIT) for transmitting beamforming and receiving beamforming, the SDMA may be supported without obtaining additional channel status information at receiver (CSIR) when realizing the uplink receiving beamforming.

Disclosed are an adaptive beamforming apparatus and method that despreads an input signal, and determines whether a symbol of despread signal belongs to a pilot sub-channel or non-pilot sub-channel of the despread signal. One of two beamforming algorithms is accordingly enabled. If the symbol belongs to the pilot sub-channel, a first algorithm is used to calculate a weight vector, and if the symbol belongs to the non-pilot sub-channel, a second algorithm is used to calculate the weight vector. A current weight vector is updated using newly calculated weight vector, and a beam pattern is formed based on the updated weight vector.

The invention belongs to the field of signal processing. The invention provides an improved Capon beamforming algorithm for the problem that the performance of a conventional Capon beamforming algorithm is significantly degraded, wherein the problem is caused by estimation errors of an array steering vector or inaccurate estimation of a data covariance matrix. On the basis that a reception data covariance matrix is reconstructed based on linear contraction estimation, a mismatch steering vector is projected to an updated subspace by updating the signal and interfering subspace, so as to acquire the optimal reception right. The algorithm maintains a high gain in the true wave incoming direction and assumed wave incoming direction of a desired signal, which avoids signal cancellation. Compared with other algorithms, the algorithm can more accurately point to the true direction of the desired signal. The algorithm is robust for steering vector mismatch caused by the observation error of the signal arrival direction.

The invention provides a real-time frequency domain super-resolution direction estimation method and a device, and the method comprises the following steps: 1) dividing a linear array for obtaining M1sub-arrays; 2) respectively carrying out beamforming on a time-space two-dimensional signal of each sub-array for obtaining an output beam of each sub-array on the scanning direction; 3) carrying outsynthesis treatment on the output beams of various sub-arrays on the scanning direction for obtaining a multi-sub-array synthetic beam; and 4) and obtaining the target direction according to the multi-sub-array synthetic beam obtained by the step 3). The invention provides the real-time frequency domain super-resolution direction estimation device. The method and the device have the following technical effects: (1) the calculation is highly efficient, the speed is fast, the DSP engineering is convenient to realize, and the real-time treatment can be realized; (2) the method and the device areapplicable to broadband noise target direction finding; (3) the high-resolution beam can be obtained; and (4) compared with the frequency domain beamforming algorithm in the prior art, the method andthe device can obtain beam output with even narrow main lobe and even weak strength of a side lobe, thereby obtaining the target direction estimation with higher resolution.

The present disclosure relates to a voice wake-up method, an apparatus and a computer readable storage medium, and relates to the field of computer technology. The method of the present disclosure includes: beamforming a speech signal in a predetermined plurality of directions to obtain a plurality of beams; input the beam into the pre-trained keyword recognition model to obtain the probability ofthe keyword contained in the beam; according to the probability that the beam contains keywords and the signal quality of the beam, the beam pointing to the sound source direction is determined as the sound source beam; and whether to wake up the system or not is determined according to the characteristic matching results of the sound source beams at a plurality of consecutive times. The invention decouples the beamforming algorithm and the acoustic source localization algorithm without using the existing acoustic source localization method and the speech wake-up process, thereby avoiding theinfluence of the acoustic source localization accuracy on the beamforming algorithm azimuth, thereby improving the wake-up accuracy of the speech system and improving the user experience.

The invention provides a digital pre-distortion structure for a beam forming system, and a control method thereof. The control method comprises the following steps: transmitting an input original signal x(n) via an antenna array after processing the original signal via a predistorter, a DAC, an up-conversion module, a phase shifter and a power amplifier; collecting an output signal yp(n) of each power amplifier by a feedback channel in a time sharing manner; synthesizing an equivalent far field signal y(n) of a main beam direction by using a beam forming algorithm according to yp(n); performing DPD training by using an indirect learning structure or a direct learning structure depending on the y(n) and the x(n), and updating the coefficients of the predistorter; and inputting the generatedDPD signal in the system, transmitting the signal via a transmitting channel, and using the transmitted signal as a linear signal of the main beam direction. The digital pre-distortion structure provided by the invention can greatly simplify the transmitter structure, reduce the computing energy consumption, achieves the linearization of the signal of the main beam direction, and can achieve verygood nonlinear performance when the difference of the nonlinear characteristics of each power amplifier is relatively large.

The invention belongs to the field of array signal processing and mainly relates to robustness of an interference-plus-noise covariance matrix reconstruction algorithm as well as robustness of optimal Capon adaptive beamforming based on the worst-case performance of the interference-plus-noise covariance matrix reconstruction algorithm. The algorithm comprises, firstly, establishing error models of steering vectors of desired signals and interference signals through worst-case performance optimization standard; secondly, estimating the power and the steering vector of a (D-1)th interference signal through Robust Capon beamforming, and meanwhile, performing EVD (evolution-data optimization) on a sample covariance matrix to estimate array receiving Gaussian white noise power and accordingly to a reconstructed interference-plus-noise covariance matrix which takes interference signal steering vector errors into consideration. The interference-plus-noise covariance matrix reconstruction algorithm can effectively solve the inherent shortcomings of summation type interference-plus-noise covariance matrix reconstruction and improve the robustness of beamforming algorithms.

A multi-beam sounding sonar water body imaging beamforming algorithm relates to the field of signal processing. The multi-beam sounding sonar water body imaging beamforming algorithm mainly comprises:compensating for the propagation loss of the acoustic wave according to a time gain curve in each detection sampling time, and obtaining a background noise level of the current detected water area after time averaging; performing near-field focusing beamforming on the signal and estimating the number of sources under the current snapshot sequence number according to the current background noise level; performing covariance matrix estimation on a signal vector with a snapshot number of 1, and obtaining a new pseudo covariance matrix by reconstructing the data covariance matrix after the forward and backward smoothness; performing singular value decomposition on the pseudo-covariance matrix, using conventional beamforming output results and an array manifold to construct a spatial spectralfunction, and obtaining a multi-beam sounding sonar water body imaging result. The algorithm can be widely applied to the multi-beam sounding sonar water body imaging function, can effectively suppress the background noise of the multi-beam sounding sonar water body imaging, and can improve the sonar imaging quality.

The invention belongs to the field of millimeter wave massive MIMO systems, in particular to a hybrid beamforming algorithm for massive MIMO. This new algorithm is designed to jointly optimize analogdomain beamforming and digital domain beamforming to maximize spectral efficiency. The invention is divided into two parts: an algorithm of analog beamforming; and the acquisition of a digital beamformer according to the "water injection method" when an analog beamformer is given. The algorithm is applicable to different types of analog networks, including continuously adjustable phase shifter networks, finite-bit adjustable phase shifter networks, switching networks, and the like. The simulation results show that the performance of the algorithm is close to the performance of the optimal all-digital beamforming, and the hybrid beamforming based on the switching networks is close to the performance of the phase shifter networks, which is more conducive to engineering implementation.

The invention discloses a double-focusing beamforming method based on self-adaptive weighting. An amplitude apodization technology, a virtual array element technology and a self-adaptive weighting method are introduced into an ultrasonic imaging system, and ultrasonic imaging of self-adaptive weighting double-focusing beamforming is realized by two-times delay superposition on the basis of a dynamic focusing technology. The double-focusing beamforming method comprises the following steps of: firstly, extracting effective information of a plurality of superposed sound fields by using the virtual array element technology; secondly, performing fixed-point focusing after performing amplitude apodization on the extracted information during first focusing; and finally, generating self-adaptive weighting coefficients by using a self-adaptive beamforming algorithm according to beamforming signals of the first focusing, and performing self-adaptive weighting beamforming focusing on the beamforming signals of the first focusing again so as to obtain data of finally imaged scanning lines. By the method, the shortcoming that a near-field region has severe pseudomorphism when the traditional double-focusing method based on virtual array elements is implemented is overcome, and an ultrasonic imaging effect is remarkably improved.

The invention discloses a multi-group and multicast joint beamforming algorithm design based on an intelligent reflecting surface. Particularly, under the single-mode constraint condition that the maximum transmitting power of the base station and the phase deviation of the IRS reflection unit are given, the method aims at designing a joint optimization strategy and algorithm of base station transmitting beams and IRS reflection phase deviation with fair user service quality. Aiming at the established non-convex and multivariable fractional optimization problem, the method is based on a fractional programming theory and an alternating optimization technology, and adopts a generalized Dinkelbach's algorithm (GDA) to carry out conversion solution. For complex sub-problems in the GDA algorithm, a matrix lifting technology is adopted to be converted into a positive semidefinite programming problem for optimization, and a feasible solution meeting the requirement that the rank is 1 is obtained through a Gaussian randomization technology. Through computer simulation verification and comparison, it is shown that the scheme provided by the invention can significantly improve the received signal to interference plus noise ratio (SINR) of the user with the worst link quality in the network.

The present invention belongs to the technical field of communication, and especially relates to an angle domain sparse beam forming algorithm based on a 3D-MIMO system noisy channel coefficient. The angle domain sparse beam forming algorithm enables a 3D beam forming problem to be modeled as an optimization model of minimum base station transmitting power by aiming at a single cell MU-MIMO downlink channel system; then converts the original problem into an equivalence optimization problem in the angle domain through angle domain conversion; and adds an l1 norm penalty term of an angle domain pre-code to the beam forming optimization problem by utilizing the angle domain sparse characteristics of a 3D-MIMO channel, so that the obtained angle domain pre-code meets certain sparsity, and is further approximate to the actual solution of the optimal angle domain pre-code, thereby improving system performance; and finally converts the optimization problem into a second-order cone programming problem through variable substitution, and then solves the problem through a convex optimization tool. A simulation result shows that the system performance of the angle domain sparse beam forming algorithm of the present invention is greatly better than the system performance of a non-robustness beam forming algorithm in the scene with big channel noise errors.

Embodiments of the invention disclose a signal separation method and device, equipment and a storage medium. The method is applied in a multi-sound source scene and comprises the steps of: carrying out uniform segmentation on a circumference formed by a ring-shaped microphone array according to angle information in a beam forming algorithm, so as to obtain beam sound areas; acquiring a mixed original sound source signal by adoption of a voice acquisition system, wherein the voice acquisition system is formed by the ring-shaped microphone array; preliminarily separating the mixed original soundsource signal to determine a preliminary sound source signal output by each beam sound area; and cancelling each preliminary sound source signal according to an echo cancellation processing so as toobtain a target sound source signal corresponding to each preliminary sound source signal. According to the method, each target sound source signal can be correctly separated; and the method can be suitable for multi-sound source scenes to provide a new thought for correctly separating each sound source in the multi-sound source scenes.

The invention provides a three-dimensional passive imaging method and system for brain focusing ultrasonic cavitation real-time monitoring. On the basis of calibrating cavitation signal distortion due to skull shielding, array elements of a surface array in two directions are subjected to robust Capon beamforming so as to greatly inhibit the interference from other directions and the mutual action between cavitation microbubbles; on the basis of the differences among phases of cavitation signals, a phase correlation coefficient is introduced for correcting a beamforming algorithm, so that the imaging resolution ratio is improved during the imaging artifact inhibition; and finally, a three-dimensional cavitation body is subjected to processing such as data thresholding and smoothening. The method and the system can solve the problems that the detection sensitivity of a conventional magnetic resonance and active ultrasonic imaging monitoring method on the cavitation is insufficient; the real-time monitoring cannot be realized by the active ultrasonic imaging monitoring; and the performance of a conventional passive imaging method is limited, and the like. A real-time monitoring measure for the cavitation effect in the brain focusing ultrasonic treatment process is provided for the clinics, so that the real-time feedback and control of the grain treatment becomes possible.

The invention provides a root value MUSIC-based mixed structure-based angle of arrival estimation method in a large-scale MIMO system. According to the method, analog beamforming and digital beamforming are combined, analog beamforming design is performed by using a phase shift network, a digital beamforming vector is designed by using a root value MUSIC algorithm, so that an angle of arrival of an arrival wave signal is estimated by using a direction ensuring highest receiving power. As the antenna array grows largely, in digital beamformign, each antenna needs a dedicated radio frequency (RF) link, while, mixed beamforming can remarkably reduce RF links so as to remarkably lower hardware cost. Compared with analog beamforming, mixed beamforming may improve the performance remarkably dueto introduction of digital beamforming. The mixed beamforming algorithm considers both the large-scale MIMO system performance and hardware cost, estimates the angle of arrival of the arrival wave signal effectively, and makes angle estimation less complex.

Audio beamforming is a technique in which sounds received from two or more microphones are combined to isolate a sound from background noise. A variety of audio beamforming spatial patterns exist. The patterns can be fixed or adapted over time, and can even vary by frequency. The different patterns can achieve varying levels of success for different types of sounds. To improve the performance of audio beamforming, a system can select a mode beam pattern based on a detected running application and / or device settings. The system can use the mode beam pattern to configure an audio beamforming algorithm. The configured audio beamforming algorithm can be used to generate processed the audio data from multiple audio signals. The system can then send processed audio data to the running application.

The invention belongs to the field of multi-beam receiving array calibration, and in particular relates to a wide coverage multi-beam receiving array calibration method. A multi-beam sounding sonar receiver circuit and a transducer array are used as the whole to measure the matrix directivity. After the system is fully warmed up, the influence of the temperature drift of the circuit is eliminated,so that the system is in a daily working state, and it is not necessary to separately measure characteristics of each extension. The measurement process adopts an automatic one-way non-stop rotationmethod, and the data acquisition is controlled by a signal synchronization line, the sampling time is accurately benchmarked, and the mechanical rotation idle running error is excluded. The array calibration is performed by using a near-field focused beamforming algorithm, and an array beam directivity curve and a beam angle error curve are calculated in a small-sized muffler pool. The response capability of the wide-coverage multi-beam receiving array to echoes at various angles can be effectively characterized, the receiving transducer array and the signal conditioning circuit are integrallymeasured; and the method can accurately reflect the overall signal response capability of the receiving system, and can be widely used in the field of multi-beam receiving array calibration.

The invention belongs to the technical field of mobile communication, in particular to a multicast transmit beamforming method and system based on angle information. In the invention, the angle information of target mobile stations is utilized to replace the complete channel information used in the traditional beamforming algorithm, thus the method and system are more suitable for a trunking communication system unable to obtain the channel information timely; and the beamforming gain is also utilized to replace the signal to noise ratio (SNR) used in the traditional beamforming algorithm to serve as the new beamforming optimization target, so the specific channel information can not be involved. The flow related to the invention comprises the following three steps: firstly, obtaining the angle information of the mobile stations by a base station; secondly, generating a beamforming vector based on that the beamforming gain serves as the optimization target; and finally, combining the beamforming vector with the signal to finish transmitting. According the invention, the received SNRs of the target users in the group sending services in trunking communication can be obviously improved; and the method and system provided by the invention are very applicable to the trunking communication system.

The method comprises only compressing and randomly sampling second harmonic components of contrast echoes received by probe array elements of ultrasonic arrays, and rejecting redundancy data by a large margin to reduce beamforming operands; further optimizing an adaptive beamforming algorithm by the compressed sensing theory, and speeding up the beamforming algorithm on the premise of no reduction of image resolution; and alternately emitting a pair of plane waves being opposite in phases by an transducer combining with a pulse inverse harmonic contrast imaging mode. A plane wave contrast image can be quickly synthesized only by twice plane wave emission while non-linear harmonic components of micro bubbles can be detected at high sensitivity and the contrast-to-tissue ratio can be greatly raised, destroys to the micro bubbles caused by multi-time radiation can be prevented, and the transient distribution of the micro bubbles can be accurately captured.

An embodiment of the invention discloses a voice signal separating method, a device thereof, electronic equipment and a storage medium. The method comprises the steps of acquiring the voice signal bymeans of a voice acquisition system, wherein the voice acquisition system comprises at least one pair of double-microphone unit which is composed of two microphones, wherein the position of each double-microphone unit is determined according to the position of a corresponding voice generating point; and separating the acquired voice signal by means of a wave beam forming algorithm, thereby obtaining a target voice source signal. Through the voice signal separating method, the device thereof, the electronic equipment and the storage medium, a target voice source signal can be accurately separated from the voice signal.

The invention discloses an audio enhancement method. Spatial spectrum of the original multi-channel audio frequency is obtained through a direction of arrival (DOA) estimation algorithm. A plurality of peak values greater than a set threshold value are obtained from the spatial spectrum; and a plurality of estimation direction values of the plurality of peak values are obtained according to the DOA estimation method. A spatial covariance matrix of the plurality of estimation direction values is obtained according to the plurality of estimation direction values and a guiding vector of a microphone array. A CGMM (complex Gaussian mixture model) is initialized and established according to the spatial covariance matrix; and a parameter of the CGMM is updated iteratively by a clustering method.The original multi-channel audio frequency is enhanced by an MVDR (minimum variance distortionless response) beamforming algorithm to obtain an enhanced audio. The method reduces the number of timesthat an EM algorithm iteratively updates the parameter of the CGMM, thereby greatly reducing the amount of calculations. Meanwhile, a category of time-frequency point masking values obtained for eachfrequency band is determined, so that masking values of the same category for each frequency band can be combined together, to solve the problem of fuzzy sequencing.

The invention belongs to the technical field of wind speed and direction measurement, and particularly relates to a wind measuring device and method based on multiple transmitting and receiving ultrasonic sensors. The device comprises an ultrasonic sensor array, a main control unit, a temperature and humidity compensation unit, an alarm unit and a display unit. According to the invention, two ultrasonic sensor subarrays composed of two ultrasonic transmitting sensors which transmit ultrasonic signals and six ultrasonic receiving sensors which receive ultrasonic signals are used to receive ultrasonic array data containing wind speed and direction information; then the array data information received by two ultrasonic sensor subarrays are merged, and the wind speed and direction informationis measured through a beamforming algorithm; different subarray flow vectors formed by two ultrasonic sensor subarrays contain more wind parameter information than an array flow vector or array flow vectors composed of two same subarray flow vectors; and the wind speed and direction measurement accuracy is higher.

In a method for selecting a transmission target terminal in a zero-forcing beamforming algorithm, information about a channel capacity of each terminal is received from all K terminals. A terminal with a largest channel capacity is selected as an initial transmission target terminal. A determination is made as to whether the number of currently selected terminals, i, is equal to the predefined number N. If the number of currently selected terminals, i, is different from the predefined number N, information about channel capacities of remaining terminals except previously selected transmission target terminals is received until i is equal to N and a terminal with a largest channel capacity is selected as a transmission target terminal.

The invention discloses a DOA estimation method based on MVDR covariance matrix element adaptive phase angle conversion. The method comprises the steps: performing Fourier transform by obtaining a valid frame for VAD detection to obtain all feature sub-frequency of a sound source signal; reconstructing a covariance matrix of the feature sub-frequency and a calculation equation of a guide vector; performing PTA decomposition and recombination on a covariance matrix element; converting the phase difference of the received data between the matrix elements into the incident angle of the sound source; optimally weighting the signal incident angle according to the microphone matrix element combination and the feature sub-frequency vector to obtain a DOA estimation of the signal source. Accordingto the method, the MVDR weight vector operation is simplified, which improves the directivity and the accuracy of the DOA estimation, and the defects that the existing beamforming algorithm engineering has low efficiency and low positioning accuracy are solved.

Methods are components are included herein for implementing beam shaping algorithms to optimize antenna beam patterns. An embodiment method includes dividing a geographical area into a plurality of geographical bins, setting up a plurality of zones for a cell based on a plurality of boundary thresholds, receiving a plurality of signal measurements from a plurality of user equipments across the geographical bins, classifying the geographical bins into the different zones by comparing the signal measurements to the boundary thresholds of the zones, calculating a plurality of gain adjustments for the corresponding geographical bins in at least some of the zones, and generating a beam shaping pattern based on the gain adjustments.

The invention belongs to the technical field of wireless communication, and discloses a 6G RIS-assisted physical layer security joint optimization method in a NOMA system. The method comprises the steps: in a downlink MISO communication network with one base station BS and two users U1 (a near-end user) and U2 (a far-end user), for the two users are located in a communication dead zone of the basestation, using RIS for auxiliary transmission, enabling the demodulation sequence of a user U1 to be prior to U2, and obtaining demodulation signal-to-noise ratios of the user ends U1 and U2; defining a security rate of the system; constructing a joint optimization equation; and dividing the joint optimization equation into two steps of beam forming and power distribution to be solved. A system comprises a power distribution module; a security rate definition module; a joint optimization equation construction module; and a joint optimization equation solving module. On the basis of combination of the RIS technology and NOMA, a joint optimization equation is established, a joint beam forming algorithm and a power distribution scheme are provided to solve the equation, and the safety rate of the system is greatly improved.

A computing device includes a first housing and a second housing attached by a hinge. The first housing includes a first microphone and the second housing includes a second microphone. After determining that an angle between the first and second housing has changed to a current angle, the computing device may determine a distance between the first microphone and the second microphone based on the current angle. A first audio signal from the first microphone and a second audio signal from the second microphone may each be modified (e.g., using a beamforming algorithm) to create first and second modified audio signals. The first and second modified audio signals may include less noise than the first and second audio signals. The first and second modified audio signals may be sent to an output jack or to an audio application.

The invention discloses a hybrid beamforming algorithm and system suitable for 5G multi-user large-scale MIMO, and the algorithm and system carry out the multi-user beamforming based on the minimum criterion of a user leakage signal and in combination with a codebook, and can eliminate the same-channel interference among users. A base station selects an appropriate vector from a user codebook through control channel broadcasting, and a user recovers a codeword vector by using the codebook after receiving the appropriate vector and takes the codeword vector as a weight of beam forming. Other users can also eliminate interference through a minimum mean square error algorithm by utilizing other received beam vectors. The algorithm is improved, the precoding matrix is quantified through the codebook, the base station can obtain the precoding matrix only by feeding back the codebook number through the user side, the information feedback amount is greatly reduced, the algorithm complexity isreduced, the system throughput is improved to some extent on the premise that extra resources are not increased, and equipment implementation is simplified.

The invention discloses a signal emission method and a corresponding receiving terminal beam forming method, and belongs to the technical field of signal processing. The emission method comprises the steps of mapping original signals into a digital signal form, inserting a comb type pilot frequency, converting signals into time-domain digital signals, performing serial and parallel conversion on the time-domain digital signals in sequence, adding a cyclic prefix, performing serial and parallel conversion and digital / analogue conversion to form analogue signals, transmitting the analogue signals through a doubly selective channel,and the like. The method applies a wave forming technology based onthe minimum mean square error to a mixed carrier system, not only realizes the natural advantage of doubly selective channel resistance of the mixed carrier system based on fractional Fourier transform, but also realizes the inhibitory effect on intersymbol interference of a wave forming algorithm, thereby greatly enhancing the transmission performance of a communication system and realizing a significant improvementonthe prior art.

Presented are a combined learning method and device using a transformed loss function and feature enhancement based on a deep neural network for speaker recognition that is robust to a noisy environment. The combined learning method using the transformed loss function and the feature enhancement based on the deep neural network for speaker recognition that is robust to the noisy environment, according to an embodiment, may comprise: a preprocessing step for learning to receive, as an input, a speech signal and remove a noise or reverberation component by using at least one of a beamforming algorithm and a dereverberation algorithm using the deep neural network; a speaker embedding step for learning to classify an utterer from the speech signal, from which a noise or reverberation component has been removed, by using a speaker embedding model based on the deep neural network; and a step for, after connecting a deep neural network model included in at least one of the beamforming algorithm and the dereverberation algorithm and the speaker embedding model, for speaker embedding, based on the deep neural network, performing combined learning by using a loss function.