36 results about "Multiple signal classification algorithm" patented technology

The invention discloses an ambient interference resisting method for testing an electromagnetic radiation emission field. The method specifically includes that a receiving antenna array is arranged in the testing field, each antenna unit can receive mixed signals composed of a tested device radiation signal and an interference signal, and by means of a spatial spectrum estimation method based on a multiple signal classification algorithm (MUSIC), the direction of arrival of each signal is obtained. According to a spatial spectrum estimation result, by means of a minimum variance distortionless response (MVDR) beam forming criterion, each antenna array element receiving signal is subjected to optimal weighting, in the premise that the tested device radiation signal is not distorted, array beams form null steering along the coming direction of the interference signal, and spatial filtering of the interference signal is achieved. According to the ambient interference resisting method, the MUSIC spatial spectrum estimation method and the MVDR beam forming technology are introduced into the field of electromagnetic compatibility testing, so that the ambient interference generated during testing of field radiation emission can be effectively inhibited, and radiation emission characteristics of a tested device in actual work environment can be accurately obtained.

The invention provides a compressed spectrum sensing method based on autocorrelation matrix reconstitution and mainly solves the problem that an existing sensing algorithm is high in sampling speed and large computation overhead. The compressed spectrum sensing method comprises that a secondary user obtains an observation sequence of a frequency spectrum environment through compressed sensing and obtains an autocorrelation matrix estimated value of Nyquist sampling by utilizing autocorrelation vector quantity of a autocorrelation matrix reconstitution Nyquist sample sequence of the observation sequence; a multi-signal classification MUSIC algorithm is adopted to obtain an estimated value of occupied channel number according to a characteristic value of the autocorrelation matrix estimated value; a characteristic spectrum is constructed according to the characteristic value and the estimated value of the occupied channel number, spectral amplitude values corresponding to characteristics of channels are added to obtain a sum, and mark numbers of the occupied channels are judged. By means of the compressed spectrum sensing method, the sampling speed of a secondary user receiving machine can be reduced, algorithm complexity at the reconstitution end is low, and spectral amplitude occupying situation in a cognitive radio system can be judged quickly.

The invention discloses a speech enhancement method. The speech enhancement method comprises: adopting a training method of a joint dictionary learning method to perform learning to obtain a joint dictionary for clean speech signals, interference and noise; measuring an incident direction of a desired speech signal by using a multiple-signal classification algorithm, adopting an MVDR beamformer toprocess signals received by an array, and enhancing the speech signal in a desired speech direction while suppressing the interference and noise in the undesired speech direction; using the joint dictionary obtained by learning to perform first-order joint sparse representation on an output signal of the MVDR beamformer, designing weight coefficients according to residual components of the sparserepresentation, obtaining the final estimated speech signal, interference and noise, obtaining a mean square meaning based optimal estimation of the speech signal in a form of a Wiener filter, finally restoring the desired speech signal, and realizing speech enhancement. The method can still maintain good speech enhancement performance in the presence of both directional non-stationary interference and non-directional environmental noise.

The invention discloses a spectrum perception method for detecting the angle of arrival of an authorized user in cognitive radio, and relates to a spectrum perception method in a cognitive radio system, in particular to a method capable of detecting the signal and the angle of arrival of the authorized user. The method comprises the following steps that: a linear array antenna system receives the signal; the spatial spectrum of the signal is obtained through multiple signal classification algorithm; the statistical average of the spatial spectrum Lambda serves as judgment reference; the threshold Gamma which meets false alarm probability Pf is obtained through a test; Lambda and threshold Gamma are compared to judge if the authorized user exists; and if the authorized user exists, the angle of arrival of the authorized user is obtained through minimum optimization searching. The method for detecting the multi-dimensional information of the authorized use provides conditions for further flexibly applying the frequency band of the authorized user.

The invention discloses a two-dimensional MUSIC (multiple signal classification) direction-detecting device for a uniform circular array, which comprises an FPGA (field programmable gate array) processing unit and a DSP (digital signal processor) processing unit, wherein the DSP processing unit is used for receiving a covariance matrix containing target azimuth information sent by the FPGA processing unit, utilizing a two-dimensional MUSIC algorithm to obtain the target azimuth information and sending the target azimuth information to the FPGA processing unit, and the FPGA processing unit is used for performing filtering noise-removing treatment on a signal received by an external antenna, and then calculating the covariance matrix containing the target azimuth information and storing the azimuth information calculated by DSP. According to the device, the two-dimensional MUSIC algorithm serves as a method for estimating the target azimuth, the calculating complexity in each step is reduced as much as possible, and the real-time direction-detecting capacity of the device is higher than that of the traditional device.

The invention provides a frequency diverse array MIMO radar target positioning method based on a fuzzy function. The positioning method comprises the steps of building a frequency diverse array MIMO according to a frequency control transmission array of an antenna, a phased array receiving array, the equal interval of a frequency control array transmission array element, the equal interval of a phased array receiving array element and a moving target at a far field; building a signal matrix received by a frequency diverse array MIMO receiving array according to the frequency diverse array MIMO; and then solving angle dimension and distance dimension information of a radar target by using the fuzzy function and a multi-signal classification algorithm, and thus completing the positioning ofthe radar target. According to the positioning method provided by the invention, the fuzzy matrix is used for replacing the related matrix to solve a noise sub-space and a signal sub-space via featurevalue decomposition, and then the two-dimensional search is performed at the time-frequency spectrum to acquire the angle and distance dimension information of the target; by using the target positioning method, the far field target positioning performance can be greatly improved, and the problem that the frequency diverse array MIMO cannot accurately estimate the target information under the condition of relatively low signal to noise ratio is solved.

The present disclosure provides a method for determining a Doppler frequency shift of a wireless signal directly reflected by a moving object. An example method includes eliminating a random phase shift caused by non-synchronization between a wireless transmitting device and a wireless receiving device by using conjugate multiplication of the channel state information (CSI) on two antennas, thereby obtaining the complete Doppler frequency shift information from the phase information of the channel state information. The example method eliminates the effect on the Doppler frequency shift caused by frequency information of static paths in a manner of removing the mean, thereby obtaining an accurate frequency estimation of the moving object. A Multiple Signal Classification (MUSIC) algorithm may be used to estimate a frequency spectrum according to practical sampling intervals to avoid an effect on the frequency estimation accuracy caused by uneven sampling rate in a practical wireless transceiving system.

The invention proposes a three-dimensional wind speed and direction measuring method based on a multiple signal classification algorithm. The method comprises the following steps: arranging ultrasonicsensors in a certain manner, and establishing an appropriate mathematical relationship between the speed, pitch angle and azimuth information of wind in a three-dimensional space; calculating time delay information when each sensor receives the information; and using the multiple signal classification algorithm to carry out subspace classification on a covariance matrix of information data received by a sensor array, calculating a spectral function and searching for a spectral peak value to obtain information about the speed, azimuth and pitch angle of the wind, so that the wind in the three-dimensional space can be accurately measured. The three-dimensional wind speed and direction measuring method based on the multiple signal classification algorithm provided by the invention does not rely on a time difference that an actual signal arrives at each sensor in the case of downwind and headwind, but uses a mathematical relationship between a wind vector and each sensor to calculate thedelay of an emission signal arriving at each receiving sensor. The method is simple to implement, does not require timing, and is less influenced by human and environmental factors.

The invention relates to a human body motion classification method based on compression perception, comprising the four steps of space-time interest point detection, video characteristic expression based on a bag-of-word model, construction of a visual dictionary and a motion classification algorithm based on compression perception. In step 1, solving training sample characteristics to obtain a training sample matrix A=[A1,A2,...AK] belong to Rm*n, k categories, a test sample y belong to RM and an optional fault tolerance degree epsilon>0; in step 2, solving a dictionary Z, a classifier W and a coefficient matrix A; and for a new video motion sequence, employing the classifier W obtained in the second step for classification, and finally obtaining the category estimation of video motion. The human body motion classification method fuses space-time interest detection, dictionary learning and video expression characteristics in a learning framework, and simultaneously learns a linear classifier; the human body motion classification method simultaneously learns a discrimination dictionary, a discrimination coding coefficient and a classifier through an optimal method, is simple to calculate, has good robustness, and enhances the capability of processing non-linear data through a compression perception method.

The present invention belongs to the radar signal processing technology field, and discloses a direction of arrival estimation method based on multiple signal classification algorithm vector correlation. The method comprises the steps of setting a radar uniform linear array, obtaining the radar reception data according to the radar uniform linear array, and determining a first steering vector and a second steering vector according to the radar uniform linear array; according to the radar reception data, calculating a covariance matrix of the radar reception data, and carrying out the eigenvalue decomposition on the covariance matrix of the radar reception data to obtain a noise subspace of the radar reception data; according to the first and second steering vectors and the noise subspace of the radar reception data, determining a first correlation vector and a second correlation vector; according to the first and second correlation vectors, constructing a spatial spectrum function; according to the spatial spectrum function, carrying out the maximum likelihood estimation on a direction of arrival to obtain an estimation value of the direction of arrival, thereby improving the angle resolution and the robustness of a direction finding performance.

The invention provides an adaptive robust zero space projection beam forming method in secure wireless transmission. The method is composed of three parts: estimation of an arrival angle and a signalto noise ratio, prediction of an angle error range and robust and secure beam forming. In a first time slot, firstly, direction angles of a desired direction and an eavesdropping direction are estimated by a root value multiple signal classification algorithm. Then, singular value decomposition is performed on a covariance matrix of a received signal. Finally, with the aid of the Kcramer's lower boundary formula, the angular error range can be predicted from the estimated arrival angle and the signal to noise ratio. In a second time slot, all the estimated parameters are used for designing a useful signal beam forming vector and a artificial noise projection vector. Compared with the traditional non-adaptive robust method and the non-robust method, the adaptive robust zero space projectionbeam forming method provided by the invention has the advantages of reducing the bit error rate and improving the security rate, so that the system realizes secure wireless transmission.

The invention discloses an automatic signal modulation and classification method and system based on K-means. Modulation characteristics of the radar and communication signals are extracted by using signal processing algorithms such as time-frequency analysis, instantaneous autocorrelation and Fourier transform, and layer-by-layer classification of the radar and communication signals of differentmodulation types is realized based on a naive Bayes and K-means machine learning algorithm. Simulation results show that the multi-signal classification algorithm network can effectively realize signal classification of six different modulation types.

The invention discloses an array scanning super-resolution microscopic imaging device, method and equipment based on a multiple signal classification algorithm and a storage medium, belongs to the field of laser scanning microscopic imaging, and aims to solve the problems that the transverse resolution of the laser scanning microscopic imaging technology is difficult to improve and the image acquisition rate is low. The system comprises a beam expanding system, a micro lens array, a collimating lens, a scanning galvanometer, a scanning lens, a tube lens, a dichroscope, an objective lens, a sample, an objective table, a collecting lens and a CCD camera. The array scanning super-resolution microscopic imaging device is used for obtaining a low-resolution image sequence of a fluorescent sample, and the low-resolution image sequence is reconstructed based on a multiple signal classification algorithm to obtain a super-resolution image. An array point illumination mode and a multiple signal classification algorithm are combined, the axial chromatography capacity is achieved while the transverse imaging resolution is improved, the imaging rate is improved through an array scanning method, and biological fluorescence samples can be imaged in a random flickering mode.

The invention discloses a corona discharge positioning method and device, electronic equipment and a storage medium. The method comprises the steps: acquiring a corona discharge signal received by an ultrasonic sensor array; carrying out discrete Fourier transform and focus conversion on the corona discharge signal, and establishing a covariance matrix; and determining the arrival direction of the corona discharge signal by using a multi-signal classification algorithm according to the covariance matrix. Because the sound wave excited by corona discharge has a frequency of up to 300 kHz and has strong directivity, it can cause the ultrasonic sensors disposed at various locations in the plane to generate different responses. According to the embodiment of the invention, the arrival direction of the corona discharge signal is calculated by using a multi-signal classification algorithm, the resolution is high, corona discharge can be conveniently and timely positioned, and the method is more intuitive and effective.

A multipath channel side writing method of the Wi-Fi equipment based on limited bandwidth is provided, a data model matched with a multi-signal classification algorithm is designed, after CSI of a plurality of antennas is reconstructed and combined, the multi-signal classification algorithm is adopted to solve the signal arrival time according to the CSI, and energy attenuation and phase change corresponding to signal arrival time are further obtained through a least square method and a Wi-Fi channel model, thereby realizing wireless channel side writing. The method is applied to the CSI-basedarrival time estimation problem by properly modifying the traditional MUSIC algorithm, so that more accurate multi-path channel side writing is provided. CSI information provided by the Wi-Fi equipment with limited bandwidth is used for accurately describing the wireless multi-path channel in detail, so that the problem of how to obtain accurate wireless multi-path channel side writing by using the narrow bandwidth Wi-Fi equipment under condition of not increasing redundant communication burden is solved.

The invention provides a generalized covariance multi-signal classification algorithm based on a score function, and belongs to the technical field of communication. The method mainly comprises the following steps: 1, selecting proper parameters, and calculating and obtaining a score function; 2, acquiring a noise-mixed signal output by the antenna array, and constructing a vector of the signal; 3, calculating and obtaining a generalized covariance matrix based on a score function; 4, calculating and obtaining the estimation of the generalized covariance matrix based on the score function. 5,carrying out eigenvalue decomposition on the generalized covariance matrix based on the score function, carrying out calculation, and obtaining estimation of a noise subspace. 6, calculating and obtaining the estimation of the spatial spectrum according to the estimation of the noise subspace. and 7, estimating the direction of arrival of the radio signal through the spectral peak position. Experimental results show that the algorithm can accurately realize direction-of-arrival estimation of radio signals under the condition of impulse noise.

The invention relates to the technical field of signal processing, in particular to a sign data processing method and system based on a frequency-modulated continuous wave radar. The method comprises the following steps: firstly transmitting an electromagnetic wave signal to a to-be-detected human body through the frequency-modulated continuous wave radar, so as to detect a chest wall mechanical motion signal of the to-be-detected human body, and receiving a returned echo signal; sampling the echo signal, processing the sampled echo signal by using a multiple signal classification algorithm to obtain a high-resolution distance image, and extracting a respiration signal and a heartbeat signal according to the high-resolution distance image; and finally determining sign data of the to-be-detected human body according to the respiration signal and the heartbeat signal, wherein the sign data comprise respiration rate, real-time respiration rate variability, heart rate and real-time heart rate variability. The method and system can perform more convenient and efficient detection of vital signs of a to-be-detected human body, realize high-precision real-time extraction of respiration rate, heart rate, respiration rate variability, and heart rate variability, and provide accurate and clear physical sign data.

The invention discloses a laser scanning super-resolution microscopic imaging device, method and equipment based on a multi-signal classification algorithm, and a storage medium, belongs to the technical field of optical precision measurement, and aims to solve the problem that the transverse resolution of a confocal microscopic technology is difficult to improve. Comprising a laser light source, and a collimator objective, a scanning galvanometer, a telecentric scanning lens, a tube lens, a dichroscope, an objective lens, a fluorescence sample, an optical filter, a collection lens and a CCD camera are sequentially arranged in the light propagation direction of the laser light source. A CCD camera is used for collecting low-resolution image sequences generated in the random flickering process of a sample in different scanning focusing light spot illumination areas, a multi-signal classification algorithm is used for reconstructing the low-resolution image sequence of each scanning position, reconstruction results corresponding to all the scanning positions are spliced, and a high-resolution image is obtained. And the transverse resolution of the confocal microscopy technology can be effectively improved.

The invention discloses a structural impact load location method based on multiple signal classification algorithm, comprising the following steps: step 1: collecting impact response signals; step 2: conducting estimation of direction-of-arrival of impact source by utilizing the multiple signal classification algorithm; and step 3: conducting estimation of distance of the impact source according to the propogation characteristic of Lamb wave. The method in the invention is convenient and reliable, can perform real-time and online monitoring to the impact; and has high-precision impact location and strong instantaneity.

The invention provides a passive indoor positioning method based on Wi-Fi, which utilizes multi-path information to perform optimal positioning of indoor target positions. First, the channel state information data is collected and preprocessed; then, the improved multiple signal classification algorithm is used for parameter estimation, and the parameter information such as the arrival angle, signal flight time, and departure angle of multiple paths is obtained; then, using these parameters The complementary relationship between the parameters is used for target positioning; finally, the clustering algorithm is used for positioning correction to obtain the final target position. Since the multi-path information is fully utilized, the method of the present invention has a better positioning effect for typical daily scenes, especially non-line-of-sight scenes.

The invention provides an indoor multi-target 3D (three-dimensional) positioning system and an indoor multi-target 3D positioning method based on radio frequency identification (RFID) and a carrier phase, and relates to the technical field of RFID and indoor positioning. In order to solve the technical problems of timeliness and precision of indoor multi-target positioning, a positioning platform is built by means of the technical advantages of radio frequency identification (RFID), positioning targets are distinguished by means of passive tags, a positioning problem is converted into an optimization problem, and by fusing the direction finding principle of a multi-signal classification (MUSIC) algorithm, the multi-carrier phase distance measurement principle and the particle swarm optimization (PSO) algorithm, the positioning accuracy of the indoor multi-target positioning is improved, and the positioning accuracy of the indoor multi-target positioning is improved. According to the method, a joint particle swarm optimization (JPSO) algorithm is proposed, the retrieval process of direction finding and distance measurement is omitted, the dependence of the positioning precision of the PSO algorithm on the number of iterations and the number of particle swarms is reduced, and multi-target synchronous positioning is directly realized. The invention provides an effective indoor multi-target rapid high-precision positioning model, the system structure is simple, the deployment is convenient, and the positioning service can be provided in most typical indoor environments.

A multipath channel side writing method of the Wi-Fi equipment based on limited bandwidth is provided, a data model matched with a multi-signal classification algorithm is designed, after CSI of a plurality of antennas is reconstructed and combined, the multi-signal classification algorithm is adopted to solve the signal arrival time according to the CSI, and energy attenuation and phase change corresponding to signal arrival time are further obtained through a least square method and a Wi-Fi channel model, thereby realizing wireless channel side writing. The method is applied to the CSI-basedarrival time estimation problem by properly modifying the traditional MUSIC algorithm, so that more accurate multi-path channel side writing is provided. CSI information provided by the Wi-Fi equipment with limited bandwidth is used for accurately describing the wireless multi-path channel in detail, so that the problem of how to obtain accurate wireless multi-path channel side writing by using the narrow bandwidth Wi-Fi equipment under condition of not increasing redundant communication burden is solved.