Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
Hiro

30 results about "Wavelet filter bank" patented technology

Tomographic reconstruction based on deep learning

The present approach relates to the use of machine learning and deep learning systems suitable for solving large-scale, space-variant tomographic reconstruction and / or correction problems. In certain embodiments, a tomographic transform of measured data obtained from a tomography scanner is used as an input to a neural network. In accordance with certain aspects of the present approach, the tomographic transform operation(s) is performed separate from or outside the neural network such that the result of the tomographic transform operation is instead provided as an input to the neural network. In addition, in certain embodiments, one or more layers of the neural network may be provided as wavelet filter banks.
Owner:GENERAL ELECTRIC CO

Method and apparatus for signal transmission and reception

A method and modem for communicating serial input data over a transmission link. Serial input data is partitioned into parallel data elements prior to rotation by an invertible linear mapping. Resulting frames of parallel signal elements sequentially modulate a carrier, which is then transmitted over the link. After receipt of the modulated carrier from the link, the signal is demodulated and assembled into frames of parallel signal elements which are de-rotated by an inverse linear mapping. Thresholding the result of the inverse mapping recovers the parallel data elements, which are then re-assembled into serial output data. The linear mapping employs: 1) commuting rotation matrices for convolutionally rotating data vectors into signal vectors and vice-versa; 2) filter bank polyphase rotation matrices; or 3) computationally efficient multi-rate wavelet filter banks. Transmitter pre-emphasis places most of the information in lower baseband frequencies; complimentary de-emphasis occurs in the receiver. Logarithmic amplification of the baseband signal prior to carrier modulation improves modulation gain and transmit channel noise attenuation. Coefficients of the rotation matrix of the receiver are adaptively equalized to correct for transmission path distortion. FM double-side band is employed in systems requiring minimized cost and complexity. FM single-side band is employed in systems in which bandwidth reduction is desirable. AM is also employable.
Owner:INTRATECH

Compensation for non-linear distortion in a modem receiver

A method and modem receiver for receiving serial data from a transmission link, for recovering data from the received data, and for synchronizing the receiver timing to that of the transmitter are disclosed. A receiver equalizer includes a filter and quantizer in the forward path and a feedback circuit connecting the quantizer output to the filter input. A measuring circuit is used to initialize the feedback circuit. The receiver filter, provided as a QMF wavelet filter bank, is implemented in one embodiment with a form of polyphase filtering followed by four M / 2-point FFTs. The quantizer recovers data symbols and provides a quantization error feedback signal. The measuring circuit provides an approximation of phase and amplitude of isolated wavelets.
Owner:MILLER WILLIAM J

Transmitter and method for digital multi-carrier transmission

The present invention relates to a transmitter and method employing a multi-carrier transmission method, especially utilizing real coefficient wavelet filter banks. The transmitter includes a preamble data generator, a modulator, and a ramp processor. The preamble data generator generates preamble bit data, and outputs the preamble data. The modulator modulates the preamble data, generates a plurality of subcarriers, and outputs a composite wave of the time waves of the plurality of subcarriers. Subsequently, the ramp processor performs ramp processing on the composite wave with a certain delay period from a reference position of the composite wave.
Owner:PANASONIC CORP

Rubbing acoustic emission denoise method based on empirical wavelet transform

The invention discloses a rubbing acoustic emission denoise method based on empirical wavelet transform. The method comprises the following steps that (1) an acoustic emission signal is acquired through a rubbing acoustic emission experimental device; (2) adaptive partition is performed on the acoustic emission signal according to the Fourier spectrum features; (3) a wavelet window is added after partition, and an empirical scale function and an empirical wavelet function are defined; (4) empirical wavelet transform is defined; and (5) wavelet denoising is performed on each empirical mode component fi and then reconstruction is performed based on EWT. The beneficial effects of the rubbing acoustic emission denoise method based on empirical wavelet transform are that adaptive partition is performed according to the Fourier spectrum of the acoustic emission signal, and a wavelet filter bank is constructed to extract different intrinsic mode components included in the acoustic emission signal so that less modes are decomposed and the phenomena of mode aliasing and endpoint effect can be effectively filtered; and wavelet denoising is performed on each empirical mode component, reconstruction is performed based on EWT and denoising is performed on the signal so that the denoised signal has high signal-to-noise ratio and the denoise effect is obvious.
Owner:SOUTHEAST UNIV

Method for constructing biorthogonal wavelet filter bank

InactiveCN104092447ASmall reconstruction error rateImprove denoising effectDigital technique networkStethoscopeVanishing momentsDecomposition
The invention discloses a method for constructing a biorthogonal wavelet filter bank. The method includes the following steps: determining the length N of a decomposition low-pass filter h(n) and the length M of a reconstruction low-pass filter (please see the specifications) according to the requirements of wavelet functions to be constructed, wherein the N and the M are even numbers; determining the orders of vanishing moments of dual wavelet generating functions (please see the specifications) to be constructed; calculating the biorthogonal complete reconstruction filter bank according to a formula set to obtain a set of coefficients of the biorthogonal complete reconstruction filter bank; substituting the results into a dual-scale equation to construct a set of biorthogonal wavelet bases (please see the specifications), decomposing a decomposition high-pass filter corresponding to the wavelet function (please see the specifications), and reconstructing a reconstruction high-pass filter corresponding to the wavelet function (please see the specifications). Compared with common db-series wavelet filter banks and common bior-series wavelet filter banks, heart sound signals are processed through the biorthogonal wavelet filter bank constructed with the method, and the better noise reduction effect, the accurate heart sound classified information and the smaller reconstruction error rate can be obtained.
Owner:NANJING UNIV OF POSTS & TELECOMM

Depth scattering convolution network learning method and system based on nuclear space

The invention relates to a depth scattering convolution network learning method and system based on a nuclear space. The system comprises a parameterization wavelet generation module, a multinuclear learning module, a scattering decomposition module and a supporting vector machine classification module. The parameterization wavelet generation module constructs a wavelet filter group via a randomized parameter. The multinuclear learning module carries out convolution decomposition of a scattering network on a training data set based on the above filter group, separately projects decomposition coefficients of convolution paths to a nuclear feature space, and selects an optimal convolution path by use of a multinuclear learning algorithm. The scattering decomposition module carrying out scattering decomposition on a test data set based on the optimal convolution path. The obtained decomposition coefficients are classified by the support vector machine classification module. According to the invention, classification accuracy of all types of digital images can be effectively improved.
Owner:SHANGHAI JIAO TONG UNIV

Compensation for non-linear distortion in a modem receiver

A method and modem for communicating serial input data over a transmission link are disclosed. Serial input data is partitioned into parallel data elements prior to rotation by an invertible linear mapping. Computationally efficient multi-rate wavelet filter banks are employed in a receiver for analyzing a received signal. Self-interference components are calculated in the receiver per sub-band and tap-weights for the filter banks are derived as a function over the range of self-interference values. A weighted sum is formed and subtracted from the delayed, received signal to eliminate the interference, and is removed from the recovered signal.
Owner:INTERTECH CORP

Magnetic control submerged-arc welding seam tracking signal analyzing method based on experience wavelet transformation

A magnetic control submerged-arc welding seam tracking signal analyzing method based on experience wavelet transformation mainly solves the problems that existing hardware filtering is not complete, low-frequency interference signals similar to seam tracking signals are difficult to filter out by a wavelet filtering method, and EMD unreasonable decomposition convergence conditions, over-envelope, under-envelope and the like are liable to cause modal aliasing. According to the key points of the technical scheme, a magnetic control submerged-arc welding arc sensor controls swinging of arcs so as to carry out welding seam groove scanning, a Hall sensor collects original signals and carries out hardware filtering processing on the original signals, the experience wavelet transformation signal analyzing method is utilized for carrying out low-frequency analyzing on the signals after the hardware filtering, one wavelet filter set is adaptively selected according to Fourier spectrum characteristics of the signals so as to extract different AM-FM components of the signals, then Hilbert conversion is carried out on each experience modal function, and meaningful instantaneous signal characteristics are obtained. By adopting the method, the more precise welding seam tracking signals are extracted, and the information of other low-frequency interference signals are obtained simultaneously.
Owner:XIANGTAN UNIV

Tomographic reconstruction based on deep learning

The present approach relates to the use of machine learning and deep learning systems suitable for solving large-scale, space-variant tomographic reconstruction and / or correction problems. In certain embodiments, a tomographic transform of measured data obtained from a tomography scanner is used as an input to a neural network. In accordance with certain aspects of the present approach, the tomographic transform operation(s) is performed separate from or outside the neural network such that the result of the tomographic transform operation is instead provided as an input to the neural network. In addition, in certain embodiments, one or more layers of the neural network may be provided as wavelet filter banks.
Owner:GENERAL ELECTRIC CO

Single measurement node simulation circuit fault diagnosis method

The invention discloses a single measurement node simulation circuit fault diagnosis method, comprising steps of obtaining prior sample data vectors under various fault modes, (2) calculating statistic average values of the prior samples under various fault modes, (3) decomposing the signal haar orthogonal wavelet filter set, extracting characteristic factors of the prior sample fault modes, (5) extracting characteristic factors of the fault mode to be detected, (6) calculating a correlation coefficient matrix and correlation degree measurement parameters between the characteristics factors of the fault mode to be detected and the characteristic factors of the prior sample fault modes, and (7) comparing the correlation degree measurement parameters and determining the fault mode according to the maximal correlation principle. The single measurement node simulation circuit fault diagnosis method disclosed by the invention converts single-path signals to multi-path signals under the condition that the original measurement information is not lost, extracts independent characteristic factors of the faulted mode, enables the characteristic factors of the faulted mode to reflect the changing situation of the circuit structure under various fault modes, studies correlated mode determination rule and finishes the successful classification of the circuit fault mode.
Owner:HEFEI UNIV OF TECH

While-drilling azimuth electromagnetic wave resistivity imaging feature extraction method

The invention discloses a while-drilling azimuth electromagnetic wave resistivity imaging feature extraction method, and the method comprises the following steps: selecting a proper dual-tree complexwavelet filter combination for the multi-scale decomposition of azimuth resistivity imaging; determining an optimal scale during the multi-scale decomposition, carrying out the analysis and comparisonof low-frequency band and high-frequency subbands at different scales after decomposition, and selecting the imaging at a scale with the optimal resolution for application; and carrying out the reconstruction of azimuth resistivity imaging. The method makes the most of the approximate translation invariance and direction selectivity of dual-tree complex wavelet transform, and achieves the dual-tree complex wavelet transform of the while-drilling azimuth electromagnetic wave resistivity imaging. The method reduces or eliminates the impact on the imaging from the interference noise. The methodcan achieve the enhancement of the texture or edge of imaging in some directions, and the detailed feature information of imaging is obtained. The method highlights the stratum boundary or crack edgeinformation reflected by imaging. The method keeps the general picture information and vibration information of imaging. The method is higher in adaptability, and needs less manual intervention.
Owner:CHANGZHOU INST OF TECH

Diesel engine high-frequency pressure oscillation signal extraction method based on improved wavelet packet transformation

The invention provides a diesel engine high-frequency pressure oscillation signal extraction method based on improved wavelet packet transformation. The method comprises the following steps: 1, normalization is performed on wavelet filter coefficients; 2, after the normalization is performed on the wavelet filter coefficients, the normalized filter coefficients are taken as individuals in a chicken flock, initialization of the chicken flock algorithm is achieved, and, after a cost function is built, wavelet filters are optimized according to individual fitness by utilizing the flock algorithm;3, a filter group obtained in step 2 is used as an initial value of a sequence secondary planning method to perform secondary iteration optimization, and a cost function is consistent with the step 1; and 4, a diesel engine cylinder pressure signal is subjected to three-layer wavelet packet transformation by utilizing the wavelet filter group built in steps 2 and 3, and a high-frequency pressureoscillation signal is extracted. The calculation amount of the diesel engine high-frequency pressure oscillation signal extraction method is greatly reduced compared with the calculation amount of anexisting wavelet improvement method; the wavelet filter group with a frequency domain characteristic very close to ideal cutoff can be built, and frequency confusion and amplitude distortion phenomenain the wavelet packet transformation process are greatly reduced.
Owner:HARBIN ENG UNIV

Two-channel linear phase wavelet filter bank with two stages of multiphase substructures

InactiveCN106411289AOvercomes the disadvantage of not having a linear phaseFewer multiplication operationsDigital technique networkVanishing momentsEngineering
The invention discloses a two-channel linear phase wavelet filter bank with two stages of multiphase substructures, which consists of an analysis filter bank, an extractor, a zero value interpolator and a synthesis filter bank. A multiphase structure of each of both the analysis filter bank and the synthesis filter bank consists of a two-input two-output filter module and a special two-input two-output module. According to the two-channel linear phase wavelet filter bank disclosed by the invention, all filters in the two-input two-output filter module have linear phases, and thus, for the wavelet filter bank disclosed by the invention, an operational speed is the same with that of a multiphase structure wavelet filter bank, but convolution multiplication times are reduced by nearly 50%; and the defect that the multiphase filters of the wavelet filter bank generally do not have the linear phases is overcome. According to the wavelet filter bank disclosed by the invention, the operational speed is the same with that of the multiphase structure wavelet filter bank, but the convolution multiplication times are reduced by nearly 50%, and one order of vanishing moment is inherent in structure; and the two-channel linear phase wavelet filter bank overcomes the defect that the multiphase filters of the wavelet filter bank generally do not have the linear phases, and has wide application prospect.
Owner:SOUTHWEST UNIVERSITY FOR NATIONALITIES

Electrocardiosignal R wave recognition method based on empirical wavelet transform (EWT) and structural feature extraction

The invention relates to an electrocardiosignal R wave recognition method based on empirical wavelet transform (EWT) and structural feature extraction, and belongs to the field of signal processing. Firstly, empirical wavelet transform is used for carrying out self-adaptive segmentation on an electrocardiosignal frequency spectrum, and a proper wavelet filter bank is constructed on a segmentationinterval to extract modal components with tight support. Then spectral analysis is carried out on each extracted modal component, to find out a high-frequency component corresponding to a R wave and carry out structural analysis on the high-frequency component, and thus accurate positioning of the R wave is achieved. Simulation results show that by using the algorithm, the average sensitivity of Rwave recognition of electrocardiosignals with noise can reach 99.93%, the average positive accuracy rate of the same can reach 99.99%, and the average accuracy rate of the same can reach 99.92%; andthe method has a good recognition effect and is of great significance to real-time monitoring of the electrocardiosignals.
Owner:CHONGQING UNIV OF POSTS & TELECOMM

Digital type dispersed software/hardware fault-tolerant type protection technique

The invention relates to a fault-tolerant protecting technology for digital discrete software and hardware, comprising the discrete technology and the fault-tolerant technology; wherein, the discrete technology realizes the discreteness of protecting software and hardware by distributed protection of CPU system based on double-loop network structure. The fault-tolerant technology comprises that: the hardware fault tolerance is realized by parallel processing of multi-CPU; protecting principle algorithm fault tolerance is realized by smart ANN technology; based on the different digital filters employed by power frequency parameter protecting principle and the wavelet filters employed by transient protection of transient signal; the filter is designed with a view to the dynamic fault-tolerant performance index, to improve the dynamic fault-tolerant performance of digital filtering algorithm; bad data in the data protection is detected and rejected which realizes information fault tolerance. The invention adopts the structure of discrete software and hardware, and the hardware employs fault-tolerant design, and the software allocates the functions of hardware; thus the invention has the advantages of algorithm fault tolerance and information fault tolerance, which respectively take the hardware adventure of the protection unit and the software adventure of the protection unit and the reliability of protection unit is improved.
Owner:GUODIAN NANJING AUTOMATION

Transmitter and method for digital multi-carrier transmission

The present invention relates to a transmitter and method employing a multi-carrier transmission method, especially utilizing real coefficient wavelet filter banks. The transmitter includes a preamble data generator, a modulator, and a ramp processor. The preamble data generator generates preamble bit data, and outputs the preamble data. The modulator modulates the preamble data, generates a plurality of subcarriers, and outputs a composite wave of the time waves of the plurality of subcarriers. Subsequently, the ramp processor performs ramp processing on the composite wave with a certain delay period from a reference position of the composite wave.
Owner:PANASONIC CORP

Unmanned aerial vehicle target rotor parameter estimation method based on scattering transformation

The invention discloses an unmanned aerial vehicle rotor parameter estimation method based on a linear frequency modulation continuous wave radar. The method comprises the steps of 1, initializing a scattering transformation scale, and generating a wavelet filter bank; 2, performing first-order scattering transformation on the echo signal; 3, setting an energy threshold value for the scattering coefficient, carrying out second-order scattering transformation, and estimating a spectral bandwidth; 4, extracting a second-order scattering coefficient where the micro-Doppler component is located, extracting a periodic component to carry out FFT (Fast Fourier Transform) to obtain a frequency spectrum, and judging the number of rotors according to the number of frequency components; 5, if the number of the rotors of the unmanned aerial vehicle is 0 determining that the rotors do not exist, and skipping to 8, otherwise, entering 6; 6, estimating the rotation frequency according to the number N of the rotors and the peak interval frequency; 7, reversely deducing the size of the blade by combining the micro-motion spectral bandwidth and the rotation frequency; and 8, outputting a rotor parameter estimation result. According to the invention, the linear frequency modulation continuous wave radar can estimate the rotating speed, the number and the size of the unmanned aerial vehicle target rotor.
Owner:HARBIN INST OF TECH

Two-channel linear phase wavelet filter bank with two-level polyphase substructure

InactiveCN102545833AOvercomes the disadvantage of not having a linear phaseDigital technique networkSubstructureMulti phase
The invention discloses an even-length two-channel linear phase wavelet filter bank with a two-level multi-phase substructure. In an analysis filter bank of the even-length two-channel linear phase wavelet filter bank, a polyphase substructure consisting of filters T0(z<2>), L1(z<2>), T1(z<2>) and L0(z<2>) is concatenated on a right side, and the other polyphase substructure is a special polyphase substructure of which all T0(z<2>), L0(z<2>) and L1(z<2>) are 1 and T1(z<2>) is -1, and is concatenated on a left side. In a comprehensive filter bank of the even-length two-channel linear phase wavelet filter bank, a polyphase substructure consisting of filters T1(z<2>), -L1(z<2>), T0(z<2>) and -L0(z<2>) is concatenated on the left side, and a polyphase substructure the same as the special polyphase substructure of the analysis filter bank is concatenated on the right side. The operating rate of the even-length two-channel linear phase wavelet filter bank can be the same as that of a wavelet filter bank with a polyphase structure, the number of multiplication operation times is about halved, and the shortcoming that a polyphase filter generally does not have a linear phase is overcome.
Owner:SOUTHWEST UNIVERSITY FOR NATIONALITIES

Even-length two-channel linear phase wavelet filter bank

InactiveCN102545832AOvercomes the disadvantage of not having a linear phaseFewer multiplication operationsDigital technique networkEngineeringMulti phase
The invention protects an even-length two-channel linear phase wavelet filter bank. According to an analysis filter bank of the even-length two-channel linear phase wavelet filter bank, a two-input and two-output filter module consists of filters T0(z<2>), L1(z<2>), T1(z<2>) and L0(z<2>); filters (1+z<-1>) and (1-z<-1>) are arranged between the input end [1] of the filter module and the input end [25] of the wavelet filter bank as well as between the input end [2] of the filter module and the input end [25] of the wavelet filter bank; according to a comprehensive filter bank, a two-input and two-output filter module consists of filters T1(z<2>),-L1(z<2>), T0(z<2>) and -L0(z<2>); and filters (1+z<-1>) and (-1+z<-1>) are arranged between the output end [13] of the filter module and the output end [26] of the wavelet filter bank as well as between the output end [14] of the filter module and the output end [26] of the wavelet filter bank. The running rate of the wavelet filter bank approximates to that of a wavelet filter bank with a multi-phase structure, but the multiplying operation frequency is reduced by nearly a half; and the shortcoming that the multi-phase filter generally has no linear phase is overcome.
Owner:SOUTHWEST UNIVERSITY FOR NATIONALITIES

Full-optical multiwavelet orthogonal frequency divisional multiplexing (OFDM) and demultiplexing

Aspects of embodiments relate to an optical wavelet transform (WT) or inverse wavelet transform (IWT) medium for implementing optical multiwavelet orthogonal Frequency Division Multiplexing (OFDM) or optical demultiplexing. In embodiments, the optical WT / IWT medium comprises a plurality of N waveguides for receiving in parallel a corresponding number of N modulated optical input signals λ1,MOD to λN,MOD. Each waveguide comprises at least two optical resonators configured such to realize a wavelet filterbank so that the modulated optical signals λ1,MOD to λn,MOD undergo WT / IWT in a manner such to obtain optical wavelet transformed signals that can be orthogonally frequency division multiplexed or demultiplexed.
Owner:MER CELLO WIRELESS SOLUTIONS

Diesel Engine High Frequency Pressure Oscillation Signal Extraction Method Based on Improved Wavelet Packet Transform

The invention provides a method for extracting high frequency pressure oscillation signal of diesel engine based on improved wavelet packet transformation. 1: Normalize the wavelet filter coefficients; 2: After normalizing the wavelet filter coefficients, use the normalized filter coefficients as individuals in the chicken flock to realize the initialization of the chicken flock algorithm. After constructing the cost function, use The chicken group algorithm optimizes the wavelet filter according to the individual fitness; 3: The filter bank obtained in step 2 is used as the initial value of the sequence quadratic programming method to perform the second iterative optimization, and the cost function is consistent with the step; 4: Using the wavelet filter bank constructed in steps 2 and 3, the three-layer wavelet packet transform is performed on the diesel engine cylinder pressure signal to extract the high-frequency pressure oscillation signal. Compared with the existing wavelet improvement method, the calculation amount of the invention is greatly reduced; the wavelet filter bank whose frequency domain characteristic is very close to the ideal cut-off can be constructed, and the frequency confusion and amplitude distortion phenomena in the wavelet packet transformation process are greatly reduced.
Owner:HARBIN ENG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products