CN107026650A - A kind of simulation low-pass filter, analog information converter and filtering method - Google Patents

A kind of simulation low-pass filter, analog information converter and filtering method Download PDF

Info

Publication number
CN107026650A
CN107026650A CN201610067654.5A CN201610067654A CN107026650A CN 107026650 A CN107026650 A CN 107026650A CN 201610067654 A CN201610067654 A CN 201610067654A CN 107026650 A CN107026650 A CN 107026650A
Authority
CN
China
Prior art keywords
signal
pass filter
simulation low
matrix
delay
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201610067654.5A
Other languages
Chinese (zh)
Other versions
CN107026650B (en
Inventor
任广梅
韩伟
刘坚能
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to CN201610067654.5A priority Critical patent/CN107026650B/en
Publication of CN107026650A publication Critical patent/CN107026650A/en
Application granted granted Critical
Publication of CN107026650B publication Critical patent/CN107026650B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • HELECTRICITY
    • H03ELECTRONIC CIRCUITRY
    • H03MCODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M1/00Analogue/digital conversion; Digital/analogue conversion
    • H03M1/66Digital/analogue converters
    • HELECTRICITY
    • H03ELECTRONIC CIRCUITRY
    • H03HIMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
    • H03H17/00Networks using digital techniques
    • H03H17/02Frequency selective networks
    • H03H17/0283Filters characterised by the filter structure

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computer Hardware Design (AREA)
  • Mathematical Physics (AREA)
  • Theoretical Computer Science (AREA)
  • Complex Calculations (AREA)
  • Filters That Use Time-Delay Elements (AREA)

Abstract

The invention discloses the simulation low-pass filter applied to AIC, analog information converter and filtering method.Simulation low-pass filter includes adder unit, state conversion unit, the first delay cell and at least one second delay cell;Adder unit is used to the mixed frequency signal of reception and the Signal averaging from state conversion unit obtaining composite signal;First delay cell is used for the delay that the unit interval is carried out to composite signal;Second delay cell is used for the delay that the unit interval is carried out to the signal from state conversion unit;State conversion unit is used to handle the signal weighting from the first delay cell and the second delay cell using loop cycle matrix, and corresponds to input respectively to adder and the second delay cell;The parameter of loop cycle matrix meets the incoherent condition of any two row of calculation matrix;One of them in first delay cell or at least one second delay cell, is additionally operable to being input to the signal after delay into the sampler being connected with simulation low-pass filter.

Description

A kind of simulation low-pass filter, analog information converter and filtering method
Technical field
The present invention relates to compressed sensing field, more particularly to a kind of simulation low-pass filter, analog information converter with And filtering method.
Background technology
It is minimum in order to ensure undistorted recovery primary signal for a broadband signal according to nyquist sampling theorem Sampling rate needs to be more than twice of highest frequency in signal.That is, the highest frequency of signal is higher, to minimum sampling speed The requirement of rate is also higher, develop higher sampling rate analog-digital converter (Analog-to-Digital Convertor, ADC difficulty) is also increasing.
In order to solve this problem, scholars have found, in many practical applications, such as radar, sonar, medical detecting system Be Deng, the information rate that signal is carried it is limited, i.e., signal have it is openness, if letter can be carried having substantially no effect on signal On the premise of breath, it will be compressed with openness signal so that the highest frequency of signal is far below primary signal after compression Highest frequency, can thus reduce the requirement to ADC sampling rate.After sampled signal is obtained, it is possible to use this is adopted The information that sample signal is carried recreates primary signal, to meet the demand using primary signal.Compressed sensing (Compressive Sensing, CS) theory is exactly to be arisen at the historic moment according to such hypothesis.Compressive sensing theory points out, if Signal is compressible or is sparse in some transform domain, then can be incited somebody to action by one with the incoherent observing matrix of transform domain High dimensional signal is projected in lower dimensional space, then high by the reconstruct of solving-optimizing problem high probability from a small amount of low-dimensional data Dimensional signal.And this utilization compressed sensing principle converts analog signals into data signal, and therefrom extract the device of sparse information Part is referred to as analog information converter (Analog-to-Information Convertor, AIC).Analog information converter one As include frequency mixer, simulation low-pass filter and sampling unit, wherein, frequency mixer is used for initial signal and pseudo-random sequence (Pseudo-Noise sequence) signal is mixed, and the low frequency part in mixed signal, which is contained, can express original The important information of signal;Simulation low-pass filter is used to be filtered the signal after mixing, by the low of the mixed signal Frequency extracting section comes out;Sampler is used to sample to filtered signal, and the sampling unit is generally ADC.
In order to realize the compressed sensing to signal, it is necessary to meet the incoherent condition of any two row of the observing matrix, The design of the condition and simulation low-pass filter is closely bound up, and the design of simulation low-pass filter is unreasonable, be able to not will meet The condition, perception is compressed to signal so as to can not realize.In summary, prior art needs one kind to realize compression sense The simulation low-pass filter known.
The content of the invention
In order to solve the above-mentioned technical problem, AIC analog low-pass filter can be applied to the embodiments of the invention provide a kind of Ripple device, a kind of it can realize the analog information converter and a kind of filtering method of compressed sensing.
In a first aspect, the embodiments of the invention provide a kind of simulation low-pass filter, the simulation low-pass filter includes:
Adder unit, state conversion unit, first delay cell and at least one second delay cell;
Wherein, the output end of the state conversion unit is connected with the adder unit, the state conversion unit it is defeated Enter end to be connected with first delay cell, the adder unit and first delay cell connection;
The output end and input of the state conversion unit are connected with least one described second delay cell respectively;
The adder unit, for the mixed frequency signal for receiving the simulation low-pass filter and from the State Transferring The signal of unit, which is overlapped, obtains composite signal, and the composite signal is input into first delay cell, described mixed Frequency signal is the signal after initial signal is mixed with random sequence signal;
First delay cell, the delay for carrying out the unit interval to the composite signal from the adder unit, And the signal after first delay units delay is input to the state conversion unit;
At least one described second delay cell, for carrying out the unit interval to the signal from the state conversion unit Delay, and the signal after the delay by least one second delay cell is input to the state conversion unit;
The state conversion unit, for using loop cycle matrix to from first delay cell and it is described at least The signal of one the second delay cell is weighted processing, and the signal after weighting corresponded to respectively input to the adder and At least one described second delay cell;The parameter of the loop cycle matrix is rushed according to the frequency domain of the simulation low-pass filter The coefficient for swashing receptance function is obtained, and the coefficient of the frequency domain impulse response function meets following condition:According to the frequency domain impulse Any two row for the calculation matrix that receptance function and sampling function are obtained are uncorrelated;
One of them second delay cell in first delay cell, or at least one described second delay cell, It is additionally operable to being input to the signal after delay into the sampler being connected with the simulation low-pass filter.
In the first possible implementation of first aspect, in the frequency domain impulse response function H at j momentj(z) it is:
Wherein, i numbers for the moment, and T is the cycle of the loop cycle matrix, row vector or column vector that N is A (t) Number, λ 'k,i,jFor Ψ (T, 0) k-th of characteristic value, λ "k,i,jFor Ψ (i-1,0) Ψ (T, k-th of characteristic value j), λ " 'k,i,j For Ψ (i-1, k-th of characteristic value j);
Wherein A (t) is loop cycle matrix.
With reference to the first possible implementation of first aspect, in second of possible implementation, the measurement Any two row of matrix are uncorrelated specifically to be represented with equation below:
Wherein, F Φi=D (z) Hi(z), Hi(z) it is the frequency domain impulse response function at the i moment, D (z) is sampling function; ΦiFor D (z) Hi(z) function after Fourier transformation, that is, the corresponding calculation matrix of moment i, F refer to that Fourier becomes Change;FΦj=D (z) Hj(z), ΦjFor D (z) Hj(z) function after Fourier transformation, that is, the corresponding measurements of moment j Matrix, F refers to Fourier transformation;(FΦi)HFor to F ΦiConjugate transposition is sought, | | F Φi||2To seek F ΦiMould 2 norms, | | F Φj||2To seek F ΦjMould 2 norms, the δ be waviness tolerance value, the VlFor calculation matrix F ΦjIt is any two row.
With reference to the first possible implementation of first aspect, in the third possible implementation, the condition Also include:The limit of system is in unit circle.
With reference to the third possible implementation of first aspect, in the 4th kind of possible implementation, the system Limit represented in unit circle with equation below:
Wherein, | | λ 'k,i,j||2To seek λ 'k,i,jMould 2 norms.
With reference to the first possible implementation of first aspect, in the 4th kind of possible implementation, the condition Also include:System pole is more than or equal to preset value.
With reference to the 4th kind of possible implementation of first aspect, in the 5th kind of mode in the cards, the system Limit is represented more than or equal to preset value by equation below:
min{||λ'k,i,j||2}≥γ
Wherein, the γ is preset value, represents lower bound of the limit apart from origin.
With reference in a first aspect, in the 6th kind of mode in the cards, the condition also includes:The loop cycle matrix Parameter within a preset range.
Second aspect, the embodiments of the invention provide a kind of analog information converter, the analog information converter includes Any possible implementation in first to the 6th kind of frequency mixer, sampler and first aspect and first aspect Simulation low-pass filter, the frequency mixer is connected with the simulation low-pass filter, and the simulation low-pass filter is adopted with described Sample device is connected;
The frequency mixer, for initial signal to be mixed with random sequence signal, obtains mixed frequency signal;
The simulation low-pass filter, for being filtered to the mixed frequency signal, obtains filtering signal;
The sampler, for being sampled to the filtering signal.
The third aspect, the embodiments of the invention provide a kind of analog information converter, including:
Processor;
Memory;And COM1;
Wherein, the processor is used for the instruction for performing the memory storage, to perform following step in execute instruction Suddenly:
Received by the COM1 come the mixed frequency signal of self-mixing device, the mixed frequency signal is initial signal and random Signal after sequence signal mixing, the frequency mixer is connected with simulation low-pass filter;
The mixed frequency signal is overlapped with the signal after weighting and obtains composite signal;
The delay of unit interval is carried out to the composite signal, and prolonging for unit interval is carried out to the signal after the weighting Late;
The signal after processing obtains the weighting is weighted to the signal by delay using loop cycle matrix; The parameter of the loop cycle matrix is obtained according to the coefficient of the frequency domain impulse response function of the simulation low-pass filter, described The coefficient of frequency domain impulse response function meets following condition:The survey obtained according to the frequency domain impulse response function and sampling function Any two row of moment matrix are uncorrelated;
Signal after delay is input to by the sampler being connected with the simulation low-pass filter by the COM1.
In the first possible implementation of the third aspect, in the frequency domain impulse response function H at j momentj(z) it is:
Wherein, i numbers for the moment, and T is the cycle of the loop cycle matrix, row vector or column vector that N is A (t) Number, λ 'k,i,jFor Ψ (T, 0) k-th of characteristic value, λ "k,i,jFor Ψ (i-1,0) Ψ (T, k-th of characteristic value j), λ " 'k,i,j For Ψ (i-1, k-th of characteristic value j);
Wherein A (t) is loop cycle matrix.
With reference to the first possible implementation of the third aspect, in second of possible implementation, the measurement Any two row of matrix are uncorrelated specifically to be represented with equation below:
Wherein, F Φi=D (z) Hi(z), Hi(z) it is the frequency domain impulse response function at the i moment, D (z) is sampling function; ΦiFor D (z) Hi(z) function after Fourier transformation, that is, the corresponding calculation matrix of moment i, F refer to that Fourier becomes Change;FΦj=D (z) Hj(z), ΦjFor D (z) Hj(z) function after Fourier transformation, that is, the corresponding measurements of moment j Matrix, F refers to Fourier transformation;(FΦi)HFor to F ΦiConjugate transposition is sought, | | F Φi||2To seek F ΦiMould 2 norms, | | F Φj||2To seek F ΦjMould 2 norms, the δ be waviness tolerance value, the VlFor calculation matrix F ΦjIt is any two row.
In the third possible implementation of the third aspect, the condition also includes:The limit of system is in unit circle It is interior.
In the 4th kind of possible implementation of the third aspect, the condition also includes:System pole is more than or equal to Preset value.
In the 5th kind of possible implementation of the third aspect, the condition also includes:The loop cycle matrix Parameter is within a preset range.Fourth aspect, the embodiments of the invention provide a kind of filtering method, methods described includes:
Simulation low-pass filter receives the mixed frequency signal for carrying out self-mixing device, and the mixed frequency signal is initial signal and stochastic ordering Signal after column signal mixing, the frequency mixer is connected with simulation low-pass filter;
The mixed frequency signal is overlapped and obtains composite signal by the simulation low-pass filter with the signal after weighting;
The simulation low-pass filter carries out the delay of unit interval to the composite signal, and to the letter after the weighting Number carry out the unit interval delay;
The simulation low-pass filter is weighted to the signal by delay using loop cycle matrix and handled Signal to after the weighting;The parameter of the loop cycle matrix is according to the frequency domain impulse response of the simulation low-pass filter The coefficient of function is obtained, and the coefficient of the frequency domain impulse response function meets following condition:According to the frequency domain impulse response letter Any two row for the calculation matrix that number and sampling function are obtained are uncorrelated;
Signal after delay is input to the sampling being connected with the simulation low-pass filter by the simulation low-pass filter Device.
In the first possible implementation of fourth aspect, in the frequency domain impulse response function H at j momentj(z) it is:
Wherein, i numbers for the moment, and T is the cycle of the loop cycle matrix, row vector or column vector that N is A (t) Number, λ 'k,i,jFor Ψ (T, 0) k-th of characteristic value, λ "k,i,jFor Ψ (i-1,0) Ψ (T, k-th of characteristic value j), λ " 'k,i,j For Ψ (i-1, k-th of characteristic value j);
Wherein A (t) is loop cycle matrix.
With reference to the first possible implementation of fourth aspect, in second of possible implementation, the measurement Any two row of matrix are uncorrelated specifically to be represented with equation below:
Wherein, F Φi=D (z) Hi(z), Hi(z) it is the frequency domain impulse response function at the i moment, D (z) is sampling function; ΦiFor D (z) Hi(z) function after Fourier transformation, that is, the corresponding calculation matrix of moment i, F refer to that Fourier becomes Change;FΦj=D (z) Hj(z), ΦjFor D (z) Hj(z) function after Fourier transformation, that is, the corresponding measurements of moment j Matrix, F refers to Fourier transformation;(FΦi)HFor to F ΦiConjugate transposition is sought, | | F Φi||2To seek F ΦiMould 2 norms, | | F Φj||2To seek F ΦjMould 2 norms, the δ be waviness tolerance value, the VlFor calculation matrix F ΦjIt is any two row.
In the third possible implementation of fourth aspect, the condition also includes:The limit of system is in unit circle It is interior.
In the 4th kind of possible implementation of fourth aspect, the condition also includes:System pole is more than or equal to Preset value.
In the 5th kind of possible implementation of fourth aspect, the condition also includes:The loop cycle matrix Parameter is within a preset range.
5th aspect, the embodiments of the invention provide a kind of filtering method, methods described is applied to first aspect and first The first of aspect is into the possible implementation of any of which in the 6th kind, and the filtering method includes:
The mixed frequency signal for carrying out self-mixing device is received, the mixed frequency signal is after initial signal is mixed with random sequence signal Signal, the frequency mixer is connected with the simulation low-pass filter;
The mixed frequency signal is overlapped and closed by the adder unit with the signal from the state conversion unit Into signal;
First delay cell carries out the delay of unit interval to the composite signal from the adder unit;
At least one described second delay cell carries out prolonging for unit interval to the signal from the state conversion unit Late;
The state conversion unit using loop cycle matrix to from first delay cell and it is described at least one The signal of second delay cell is weighted processing, and the signal after weighting is corresponded to input respectively to the adder and described At least one second delay cell;The parameter of the loop cycle matrix is rung according to the frequency domain impulse of the simulation low-pass filter The coefficient of function is answered to obtain, the coefficient of the frequency domain impulse response function meets following condition:According to the frequency domain impulse response Any two row for the calculation matrix that function and sampling function are obtained are uncorrelated;
It is single using one of them second delay in first delay cell, or at least one described second delay cell Member, the signal after delay is input to the sampler being connected with the simulation low-pass filter.
In the first possible implementation of the 5th aspect, in the frequency domain impulse response function H at j momentj(z) it is:
Wherein, i numbers for the moment, and T is the cycle of the loop cycle matrix, row vector or column vector that N is A (t) Number, λ 'k,i,jFor Ψ (T, 0) k-th of characteristic value, λ "k,i,jFor Ψ (i-1,0) Ψ (T, k-th of characteristic value j), λ " 'k,i,j For Ψ (i-1, k-th of characteristic value j);
Wherein A (t) is loop cycle matrix.
With reference to the first possible implementation of the 5th aspect, in second of possible implementation, the measurement Any two row of matrix are uncorrelated specifically to be represented with equation below:
Wherein, F Φi=D (z) Hi(z), Hi(z) it is the frequency domain impulse response function at the i moment, D (z) is sampling function; ΦiFor D (z) Hi(z) function after Fourier transformation, that is, the corresponding calculation matrix of moment i, F refer to that Fourier becomes Change;FΦj=D (z) Hj(z), ΦjFor D (z) Hj(z) function after Fourier transformation, that is, the corresponding measurements of moment j Matrix, F refers to Fourier transformation;(FΦi)HFor to F ΦiConjugate transposition is sought, | | F Φi||2To seek F ΦiMould 2 norms, | | F Φj||2To seek F ΦjMould 2 norms, the δ be waviness tolerance value, the VlFor calculation matrix F ΦjIt is any two row.
In the third possible implementation of the 5th aspect, the condition also includes:The limit of system is in unit circle It is interior.
In the 4th kind of possible implementation of the 5th aspect, the condition also includes:System pole is more than or equal to Preset value.
In the 5th kind of possible implementation of the 5th aspect, the condition also includes:The loop cycle matrix Parameter is within a preset range.
Mixed frequency signal is inputted addition list by the simulation low-pass filter that the present invention is designed it can be seen from above-mentioned technical proposal After member, by adder unit being superimposed mixed frequency signal and the signal from state conversion unit and by by the first delay The circular treatment that unit, at least one second delay cell and state conversion unit are postponed, weighted, is realized to mixing letter Number filtering.
Wherein, the analogue low pass filtering that the parameter of loop cycle matrix is designed according to the present invention in the state conversion unit The frequency domain impulse response function that device is obtained is obtained, and the parameter for the frequency domain impulse response function that the simulation low-pass filter is obtained is needed The incoherent condition of any two row of calculation matrix is met, so that simulation low-pass filter can be applied to analog information conversion Device, the frequency domain impulse response function and sampling function that the calculation matrix is obtained according to the simulation low-pass filter is obtained.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing There is the accompanying drawing used required in technology description to be briefly described, it should be apparent that, drawings in the following description are only this Some embodiments of invention, for those of ordinary skill in the art, without having to pay creative labor, may be used also To obtain other accompanying drawings according to these accompanying drawings.
Fig. 1 is a kind of structural representation for simulation low-pass filter that the embodiment of the present invention one is provided;
Fig. 2 is the structural representation for the second order simulation low-pass filter that the embodiment of the present invention one is provided;
Fig. 3 is a kind of structural representation for analog information converter that the embodiment of the present invention two is provided;
Fig. 4 is a kind of structural representation for analog information converter that the embodiment of the present invention three is provided.
Embodiment
Before technical scheme is introduced, the basic theories of compressed sensing is introduced first.It is assumed that signal can Represented using the signal x (signal vector) using a length as n, when signal x is with one group of orthogonal basisWhen linear combination is represented, I.e.Wherein Ψ is sparse transformation basic matrix.If projection coefficient θ only has K significant coefficient (i.e. definitely Value very big coefficient relatively), and K is much smaller than n, it is openness that we just say that signal x has, and signal x is also referred to as K- sparse signals. Because signal x is sparse, therefore it can be compressed, it is assumed that the signal obtained after overcompression is y (signal vector), length is M, and m is much smaller than n, then signal y can be referred to as observation vector.Observation vector is the linear projection of signal vector, this throwing Shadow process can represent with calculation matrix Φ, i.e. y=Φ x, wherein, calculation matrix Φ is a rank deficient matrix.That is, false If projection coefficient is (θ12,…,θn), sparse transformation basic matrix Ψ isCalculation matrix Φ isSo Because m is much smaller than n, it is achieved that the purpose of compressed signal.In addition, Φ and Ψ product A is called observing matrix by we.
Signal y after compression has been obtained is, it is necessary to which restoring it can just use for signal x.In order to realize that signal recovers, It has to be ensured that the correlation of observing matrix A (A=Φ Ψ) any two row is as small as possible.Because sparse transformation basic matrix Ψ is It is relatively-stationary, therefore it is as small as possible to only need to ensure the correlation of calculation matrix Φ any two row.
In the embodiment of the present invention, the compressed sensing to primary signal is realized using analog information converter AIC, i.e., according to such as Upper described principle carries out sampling in compression to signal.Based on above-mentioned analysis, in order to realize the requirement of compressed sensing, AIC Design need to ensure that the correlations of calculation matrix Φ any two row are as small as possible.And for AIC, it generally comprises three Individual part, is frequency mixer, low-pass analog filter and sampler respectively, in this three, can be based on low-passed analog filter The low-passed analog filter function of ripple device and the sampling function of the sampler obtain the calculation matrix Φ.Due to sampling function one As it is only related to sample rate, become more readily available, therefore, in AIC design, the design of simulation low-pass filter is just particularly closed Key.It should be reached by the design to simulation low-pass filter so that uncorrelated this requirement of any two row of the calculation matrix Φ, So as to the need for meeting compressed sensing.
In the present invention, inventor designs a simulation low-pass filter first, according to each list in the simulation low-pass filter The annexation of member, determines the relational expression of the frequency domain impulse response function of the simulation low-pass filter, based on the measurement square Correlation this requirement as small as possible of any two row of battle array Φ, it is determined that the value of a variable in the relational expression is obtained, with reference to this The value of variable, it is determined that the matrix parameter of corresponding units in the simulation low-pass filter is drawn, so as to complete for institute in AIC State the design of simulation low-pass filter.Above-mentioned design is described in detail below.
First, the design of the composition of simulation low-pass filter and annexation
In the present invention, the simulation low-pass filter includes adder unit, state conversion unit, a first delay list At least one first second delay cell.The output end of the state conversion unit is connected with the adder unit, and the state turns The input for changing unit is connected with the first delay cell end, the adder unit and first delay cell connection;Institute The output end and input for stating state conversion unit are connected with least one described second delay cell respectively.The addition list Member, is overlapped with the signal from state conversion unit for the mixed frequency signal for receiving the simulation low-pass filter, obtains First delay cell is input to composite signal, and by the composite signal.First delay cell, for from The composite signal of the adder unit carries out the delay of unit interval, and the signal Jing Guo first delay units delay is defeated Enter to the state conversion unit.At least one described second delay cell, for the letter from the state conversion unit The delay of unit interval number is carried out, and the signal after the delay by least one second delay cell is input to described State conversion unit.The state conversion unit, for utilizing loop cycle matrix to from first delay cell and institute The signal for stating at least one the second delay cell is weighted processing, and the signal after weighting corresponded to respectively input to it is described extremely Few second delay cell and the adder, to be circulated.First delay cell, or it is described at least one second One of them second delay cell in delay cell, is additionally operable to the signal after delay being input to and the analogue low pass filtering The sampler of device connection.
According to the above-mentioned composition and annexation of simulation low-pass filter, frequency domain impulse response function H can be obtainedj(z)。
2nd, the parameter of the loop cycle matrix in state conversion unit determines whether simulation low-pass filter can be applied Key in AIC systems is the determination of loop cycle matrix parameter, the most important condition that the loop cycle matrix parameter is met It is exactly that any two row of calculation matrix are uncorrelated.Because filter function Hj(z) parameter is determined by the parameter of loop cycle matrix Fixed, and the calculation matrix is by frequency domain impulse response function Hj(z) determined with sampling function D (z) product, described Sampling function D (z) is known, frequency domain impulse response function Hj(z), it is known that but frequency domain impulse response function Hj(z) feelings of unknown parameters Under condition, according to frequency domain impulse response function Hj(z) matrix carried out again after Fourier transformation that is multiplied with sampling function D (z) is any Two row are uncorrelated, can obtain frequency domain impulse response function Hj(z) coefficient, then according to the frequency domain impulse response function Hj (z) the anti-parameter for releasing loop cycle matrix of coefficient.
Embodiment one:
The operation principle of simulation low-pass filter based on above-mentioned introduction, the present invention provides a kind of simulation applied to AIC The embodiment of low pass filter, referring to Fig. 1, the simulation low-pass filter includes:
Adder unit 101, state conversion unit 102, first delay cell 103 and at least one second delay cell 104。
Wherein, the output end of the state conversion unit 102 is connected with the adder unit 101, the State Transferring list The input of member 102 is connected with first delay cell 103, and the adder unit 101 and first delay cell 103 connect Connect;
The output end and input of the state conversion unit 102 respectively with least one described second delay cell 104 Connection.
The adder unit 101, for the mixed frequency signal for receiving the simulation low-pass filter and from State Transferring The signal of unit 102 is overlapped, and obtains composite signal.The mixed frequency signal comes from frequency mixer, specifically, the frequency mixer Initial signal is mixed with random sequence signal, the mixed frequency signal is obtained.
First delay cell 103, for carrying out the unit interval to the composite signal from the adder unit 101 Delay, and the signal after first delay cell 103 postpones is input to the state conversion unit 102.
At least one described second delay cell 104, it is single for being carried out to the signal from the state conversion unit 102 The delay of position time, and the signal after the delay by least one second delay cell 104 is input to the state Converting unit 102;
The state conversion unit 102, for utilizing loop cycle matrix to from first delay cell 103 and institute The signal for stating at least one the second delay cell 104 is weighted processing, and the signal after weighting is corresponded into input respectively to institute Adder 101 and at least one described second delay cell 103 are stated, to be circulated.The loop cycle matrix refers to parameter The characteristics of periodically variable circular matrix, circular matrix is that each parameter of row vector is each of previous row row vector Individual parameter moves right the result of a position successively.In the present embodiment, the loop cycle matrix can be expressed as:
The loop cycle matrix has the characteristic of mechanical periodicity, that is, A (t)=A (t+T), T are the loop cycle The cycle of matrix.When the T is 0, constant matrix when the loop cycle matrix is referred to as, it is adaptable to steady random initial letter Number, the steady random any n dimension distribution functions for being meant that signal or probability density function are unrelated with start time;Work as institute State T more than 0 when, the loop cycle matrix be time-varying matrix, it is adaptable to the random initial signal of non-stationary, the non-stationary with Machine refers to that any n dimension distribution functions or probability density function of signal are related to start time.
The parameter of the loop cycle matrix is according to the coefficient of the frequency domain impulse response function of the simulation low-pass filter Obtain, the coefficient of the frequency domain impulse response function meets following condition:According to the frequency domain impulse response function and sampling letter Any two row for the calculation matrix that number is obtained are uncorrelated.
First delay cell 103, is additionally operable to the signal after first delay cell 103 is postponed and is input to and institute State the sampler of simulation low-pass filter connection;Or, at least one described second delay cell 104 one of them second Delay cell 104, is additionally operable to the signal after second delay cell 104 is postponed and is input to and simulation low-pass filter company The sampler connect.
Below the analog low-pass filter is introduced by taking first delay cell 103 and one second delay unit 104 as an example The operation principle of ripple device.Referring to Fig. 2, it is assumed that the mixed frequency signal of t down-conversion mixers output is not x (t), the loop cycle in the same time MatrixMixed frequency signal x (t) input is to adder unit 101, by described 101 pairs of institutes of adder unit State mixed frequency signal x (t) and the signal y from state conversion unit 1021(t) synthesized, the signal after synthesis is z (t).Close Into delays of the signal z (t) Jing Guo the first delay cell 103, as signal x1(t), the signal x1(t) it is divided into two-way, it is defeated all the way Go out, another road is inputted to state conversion unit 102.Meanwhile, the signal x from the second delay cell 1042(t) shape is also input to State converting unit 102.The signal x1(t) with signal x2(t) signal vector, the cycle with the state conversion unit 102 are constituted Circular matrix A (t) is multiplied, with to signal x1(t) with signal x2(t) it is weighted, and exports the signal y after weighting1And y (t)2 (t), wherein signal y1(t) input to adder unit 101, signal y2(t) input is to the second delay cell 104, to carry out next time Circulation.
Specifically, assume that x (0), x (1), x (2) they are respectively the mixed frequency signal of frequency mixer output at moment t=0,1,2, Wherein t=0 represents initial time.Assuming that A (0), A (1), A (2) are respectively the corresponding loop cycle square at moment t=0,1,2 Battle array, wherein, z(0)、z(1)、 Z (2) is respectively the signal that adder unit 101 is exported at moment t=0,1,2;x1(0)、x1(1)、x1(2) it is respectively in moment t The signal that first delay cell 103 is exported when=0,1,2;x2(0)、x2(1)、x2(2) it is respectively at moment t=0,1,2 second The signal that delay cell 103 is exported;y1(0)、y1(1)、y1(2) be respectively at moment t=0,1,2 state conversion unit 102 to The signal that adder unit 101 is exported;y2(0)、y2(1)、y2(2) be respectively at moment t=0,1,2 state conversion unit 102 to The signal of second delay cell 104 output.
As t=0, it can obtain:
Z (0)=x (0)
x1(0)=0, x2(0)=0
y1(0)=0, y2(0)=0
It should be noted that at the t=0 moment, mixed frequency signal x (0) passes through adder unit 101, due to it is initial when y1(0)= 0, therefore the composite signal z (0) of output adder unit 101 remains as mixed frequency signal x (0).
As t=1, it can obtain:
x1(1)=z (0)=x (0), x2(1)=0
Z (1)=y1(1)+x(1)
It should be noted that at the t=1 moment, signal x (0) have passed through the first delay cell 103 and be changed into x1(1), but letter Number x1(1) it is identical with signal x (0) value.And the time that first delay cell 103 postpones, i.e. unit interval, it is exactly t=1 With t=0 time difference.Moreover, at the t=1 moment, second delay cell 104 does not have signal outflow, therefore signal x2(1) Value be 0.
As t=2, it can obtain:
x1(2)=z (1)=y1(1)+x (1)=a1(1)x(0)+x(1)
x2(2)=y2(1)=a2(1)x(0)
Z (2)=y1(2)+x (2)=(a1(2)a1(1)+a2(2)a2(1))x(0)+a1(2)x(1)+x(2)
The structure and operation principle of simulation low-pass filter are mainly described above, in order to apply in AIC systems In, realize the compression to signal, in addition it is also necessary to which the parameter in loop cycle matrix is designed.The cycle is described in detail below The calculating process of the parameter of circular matrix.
First, the low-pass analog filter designed according to the present embodiment, can obtain moment j frequency domain impulse response function Hj(z):
Wherein, i be the moment numbering, T be loop cycle matrix cycle, N for A (t) row vector or column vector number, λ'k,i,jFor Ψ (T, 0) k-th of characteristic value, λ "k,i,jFor Ψ (i-1,0) Ψ (T, k-th of characteristic value j), λ " 'k,i,jFor Ψ (i-1, k-th of characteristic value j).
Function Ψ (n2,n1) definition be:
Wherein I is that the element on unit matrix, i.e. the matrix upper left corner to the diagonal (being referred to as leading diagonal) in the lower right corner is equal For 1, other elements are all 0.
Therefore it can obtain:
Wherein A (t) is loop cycle matrix.
The physical meaning of so-called characteristic value is:Assuming that matrix A (such as Ψ (T, 0), Ψ (i-1,0) Ψ (T, j), Ψ (i-1, J)) it can be expressed as:Ax=ax, wherein a are not 0, then a is representing matrix A characteristic value, and corresponding characteristic vector is x.Can To think that matrix A has embodiment on characteristic vector x direction, and it is a in characteristic vector x embodiment value.If a is bigger, square is represented The power contributed in the x direction of battle array A is bigger, and information content is more, and be here used for solving the simulation low-pass filter is Number.
Obtaining moment j frequency domain impulse response function Hj(z) after, it is assumed that the corresponding sampling function of sampler 31 is D (z), by the frequency domain impulse response function Hj(z) obtained after being multiplied with sampling function D (z):
j=D (z) Hj(z) (2)
Wherein, ΦjFor D (z) Hj(z) function after Fourier transformation, that is, calculation matrix.F refers to that Fourier becomes Change.
As previously described, in order to realize that simulation low-pass filter can be applied to AIC systems, the compression sense of signal is realized Know, it is necessary to ensure that the correlation of any two row of calculation matrix is uncorrelated, but in actual applications for various reasons, this Part can not almost be realized, therefore ensure that the correlation of any two row of calculation matrix is as small as possible, thus have following condition:
Wherein, F ΦiFor the corresponding calculation matrix of moment i, F ΦjFor the corresponding calculation matrix of moment j, (F Φi)HFor to F ΦiConjugate transposition is sought, | | F Φi||2To seek F ΦiMould 2 norms, | | F Φj||2To seek F ΦjMould 2 norms, the δ is Waviness tolerance value, the VlFor calculation matrix F ΦjIt is any two row.That is, calculation matrix F ΦjThe correlation of any two row Property be less than or equal to default waviness tolerance value δ, you can think calculation matrix F ΦjAny two row are uncorrelated.
In summary, because inventor devises the low-pass analog filter of said structure, therefore frequency domain punching can be obtained Swash receptance function Hj(z), the frequency domain impulse response function Hj(z) parameter is the ginseng according to the loop cycle matrix A (t) What number was obtained.According to the frequency domain impulse response function Hj(z) product with sampling function D (z) can obtain calculation matrix, profit The condition of uncorrelated (i.e. formula (3)) is arranged with the calculation matrix any two, frequency domain impulse response function H can be obtainedj(z) Parameter, then can utilize frequency domain impulse response function Hj(z) parameter instead releases loop cycle matrix A (t) according to formula (1) Parameter.After the parameter for obtaining loop cycle matrix A (t), it is possible to by the low-pass analog filter realization pair of said structure The compressed sensing of signal.
In addition, in actual applications, the frequency domain impulse response function H of the low-pass analog filterj(z) parameter except The incoherent condition of any two row of calculation matrix is met, following condition should be also met:
1st, as simulation low-pass filter, in addition it is also necessary to which guarantee system is stable in itself, so-called system is stable to be sought to protect All limits of card system are all in unit circle, and the limit represents the position of simulation low-pass filter passband, the unit circle It is 0 to refer to origin, the circle for the point composition for being 1 apart from the Euclidean distance of origin.Need to meet following condition:
Wherein, | | λ 'k,i,j||2To seek λ 'k,i,jMould 2 norms, λ 'k,i,jImplication referring to formula (1).
In the present embodiment, when loop cycle matrix meets the condition of formula (3) and formula (4), it is possible to ensureing mould Intend low pass filter on the basis of its own system is stable, the simulation low-pass filter for designing the present embodiment is applied to AIC, real The compressed sensing of existing signal.
2nd, the information of more input values is carried, it is necessary to allow in observation in order to ensure higher recovery probability, therefore Need to ensure that there is simulation low-pass filter longer system to respond during design.In order to reach this purpose, the condition is also wrapped Include:System pole is more than or equal to preset value, is specifically represented with formula (5).
min{||λ'k,i,j||2}≥γ (5)
Wherein, the γ is preset value, represents lower bound of the limit apart from origin.
3rd, in order to ensure convergence of algorithm speed, in the parameter of calculating cycle circular matrix, it can be made in default model Interior value is enclosed, for example [- 5,5].
Embodiment two:
The simulation low-pass filter designed based on inventor, the present invention also provides a kind of analog information converter, joins See Fig. 3, the analog information converter includes frequency mixer 11, simulation low-pass filter 21 and sampler 31.The frequency mixer 11 Connected with the simulation low-pass filter 21, the simulation low-pass filter 21 and the sampler 31 are attached.
Initial signal is mixed by the frequency mixer 11 with random sequence signal, obtains mixed frequency signal.
The frequency mixer 11 can not do specific limit using all frequency mixers suitable for AIC in the prior art, the present invention It is fixed.Frequency mixer 11 is used to initial signal being mixed with random sequence signal, wherein, the random sequence signal is general by one Individual takes the random sequence of { ± 1 } two kinds of values to constitute, and in order to be mixed all initial signals obtained in bandwidth, the stochastic ordering The sample rate of column signal is higher than the sample rate of initial signal.The part letter of the feature of initial signal can be expressed in mixed frequency signal The low frequency part of mixed frequency signal number is just concentrated on, so passes through simulation low-pass filter 21, it is possible to by the low frequency of mixed frequency signal Extracting section comes out, and realizes the compression to initial signal.
The composition and annexation of the simulation low-pass filter 21 are referring to embodiment one, and the present embodiment is repeated no more.
The sampler 31, for being sampled to the filtering signal.
Due to the filtering signal exported from simulation low-pass filter 21, its highest frequency is less than the most high frequency of initial signal Rate, according to Nyquist's theorem, the sampling rate of the sampler 31 will be less than the sampling directly sampled to initial signal Speed, that is, reduce the requirement to the sampling rate of sampler 31.
In actual applications, the sampler 31 can use all samplers suitable for AIC systems, and the present invention is not done It is specific to limit.In the present embodiment, the sampler 31 is ADC (Analog to Digital Converter, analog-to-digital conversion Device).
Embodiment three
In order to further reduce the requirement to sampler samples rate, the AIC systems of inventor's design are included parallel at least Frequency mixer, simulation low-pass filter and the sampler mentioned in two sets of embodiments one, the mould provided referring to Fig. 4, the present embodiment Intending transcriber includes at least one frequency mixer 51, at least one simulation low-pass filter 52 and at least one sampler 53, Wherein, the quantity of these three devices is identical, and the function of this three classes device is low with the frequency mixer 11 in embodiment one, simulation respectively Bandpass filter 21 and sampler 31 are identical, and here is omitted.It should be noted that in order to ensure AIC systematic survey matrixes The correlation of any two row is as small as possible, and the random sequence signal in each frequency mixer 51 should be different, and each mould Corresponding loop cycle matrix is different in plan low pass filter 52, and the parameter of each loop cycle matrix needs each mould of simultaneous Intend the condition that should meet of low pass filter 52 to solve.
Because the random sequence signal in frequency mixer 51 is different, and corresponding week in each simulation low-pass filter 52 Phase circular matrix is also different, thus while input to AIC signal is identical, but from each simulation low-pass filter The signal of 52 outputs is different, is sampled respectively by each signal to output, it is possible to obtain different sampled datas, real Requirement of the reduction to the sample rate of sampler 53 is showed.For example, it is assumed that the sample rate for requiring sampler 53 is fs, if there is two groups to mix Frequency device, simulation low-pass filter and sampler, then the sample rate of each sampler can be just fs/2.Group number is more, to every The requirement of the sample rate of individual sampler is lower.
It should be noted that each unit in the simulation low-pass filter mentioned in the embodiment one to three can be used Software is realized, can also be realized by hardware, for example adder unit can be realized by adder, and delay cell, which can be utilized, prolongs Slow device realizes that state conversion unit can be realized using memory and adder, specifically, adder, delayer, can pass through Resistance, operational amplifier, electric capacity, diode etc. are realized.In this regard, the present invention is not specifically limited, those skilled in the art can Voluntarily to select according to the actual requirements.If realized using software, if for example using matlab instruments, (matlab is individual soft Part), frequency mixer, adder, the delayer model buildings in matlab simulink instruments can be used, respective devices are realized Function;If using matlab code, frequency mixer is just directly with " * " order, z=x*y, adder "+" order, i.e. z =x+y, delayer is with z (n)=x (n-1), to simulate the function of respective devices.
Example IV
The simulation low-pass filter provided based on embodiment one, the present invention also provides a kind of embodiment of filtering method, institute State in the simulation low-pass filter that filtering method is referred to applied to embodiment one to three, methods described includes:
The mixed frequency signal for carrying out self-mixing device is received, the mixed frequency signal is after initial signal is mixed with random sequence signal Signal, the frequency mixer is connected with the simulation low-pass filter;
The mixed frequency signal is overlapped and closed by the adder unit with the signal from the state conversion unit Into signal;
First delay cell carries out the delay of unit interval to the composite signal from the adder unit;
At least one described second delay cell carries out prolonging for unit interval to the signal from the state conversion unit Late;
The state conversion unit using loop cycle matrix to from first delay cell and it is described at least one The signal of second delay cell is weighted processing, and the signal after weighting is corresponded to input respectively to the adder and described At least one second delay cell;The parameter of the loop cycle matrix is rung according to the frequency domain impulse of the simulation low-pass filter The coefficient of function is answered to obtain, the coefficient of the frequency domain impulse response function meets following condition:According to the frequency domain impulse response Any two row for the calculation matrix that function and sampling function are obtained are uncorrelated;
It is single using one of them second delay in first delay cell, or at least one described second delay cell Member, the signal after delay is input to the sampler being connected with the simulation low-pass filter.
The operation principle of the filtering method provided on the present embodiment specifically refers to embodiment one, and here is omitted.
Embodiment five
The simulation low-pass filter provided based on embodiment one, the present invention also provides a kind of another implementation of filtering method Example, methods described includes:
Simulation low-pass filter receives the mixed frequency signal for carrying out self-mixing device, and the mixed frequency signal is initial signal and stochastic ordering Signal after column signal mixing, the frequency mixer is connected with simulation low-pass filter;
The mixed frequency signal is overlapped and obtains composite signal by the simulation low-pass filter with the signal after weighting;
The simulation low-pass filter carries out the delay of unit interval to the composite signal, and to the letter after the weighting Number carry out the unit interval delay;
The simulation low-pass filter is weighted to the signal by delay using loop cycle matrix and handled Signal to after the weighting;The parameter of the loop cycle matrix is according to the frequency domain impulse response of the simulation low-pass filter The coefficient of function is obtained, and the coefficient of the frequency domain impulse response function meets following condition:According to the frequency domain impulse response letter Any two row for the calculation matrix that number and sampling function are obtained are uncorrelated;
Signal after delay is input to the sampling being connected with the simulation low-pass filter by the simulation low-pass filter Device.
The operation principle of the filtering method provided on the present embodiment specifically refers to embodiment one, and here is omitted.
Embodiment six
The present invention also provides the embodiment of another simulation low-pass filter, and the simulation low-pass filter includes:
Processor;
Memory;And COM1;
Wherein, the processor is used for the instruction for performing the memory storage, to perform following step in execute instruction Suddenly:
Received by the COM1 come the mixed frequency signal of self-mixing device, the mixed frequency signal is initial signal and random Signal after sequence signal mixing, the frequency mixer is connected with simulation low-pass filter;
The mixed frequency signal is overlapped with the signal after weighting and obtains composite signal;
The delay of unit interval is carried out to the composite signal, and prolonging for unit interval is carried out to the signal after the weighting Late;
The signal after processing obtains the weighting is weighted to the signal by delay using loop cycle matrix; The parameter of the loop cycle matrix is obtained according to the coefficient of the frequency domain impulse response function of the simulation low-pass filter, described The coefficient of frequency domain impulse response function meets following condition:The survey obtained according to the frequency domain impulse response function and sampling function Any two row of moment matrix are uncorrelated;
Signal after delay is input to by the sampler being connected with the simulation low-pass filter by the COM1.
The operation principle of the simulation low-pass filter provided on the present embodiment specifically refers to embodiment one, herein no longer Repeat.
" first " for the first router mentioned in the embodiment of the present invention is used only to do name mark, not representative order On first.The rule is equally applicable to " second ".
It should be noted that each embodiment in this specification is described by the way of progressive, each embodiment it Between identical similar part mutually referring to what each embodiment was stressed is the difference with other embodiment. For embodiment two to six, because it is substantially similar to embodiment one, so describe fairly simple, correlation Place illustrates referring to the part of embodiment one.Device embodiment described above is only schematical, is divided wherein being used as Unit from part description can be or may not be it is physically separate, the part shown as unit can be or It may not be physical location, you can with positioned at a place, or can also be distributed on multiple NEs.Can basis Some or all of module therein is selected to realize the purpose of this embodiment scheme the need for reality.Ordinary skill people Member is without creative efforts, you can to understand and implement.
The foregoing is only a preferred embodiment of the present invention, but protection scope of the present invention be not limited thereto, Any one skilled in the art the invention discloses technical scope in, the change or replacement that can be readily occurred in, It should all be included within the scope of the present invention.Therefore, protection scope of the present invention should be with scope of the claims It is defined.

Claims (21)

1. a kind of simulation low-pass filter, it is characterised in that the simulation low-pass filter includes:
Adder unit, state conversion unit, first delay cell and at least one second delay cell;
Wherein, the output end of the state conversion unit is connected with the adder unit, the input of the state conversion unit It is connected with first delay cell, the adder unit and first delay cell connection;
The output end and input of the state conversion unit are connected with least one described second delay cell respectively;
The adder unit, for the mixed frequency signal for receiving the simulation low-pass filter and from the state conversion unit Signal be overlapped and obtain composite signal, and the composite signal is input to first delay cell, the mixing letter Signal after number being mixed for initial signal and random sequence signal;
First delay cell, the delay for carrying out the unit interval to the composite signal from the adder unit, and will Signal after first delay units delay is input to the state conversion unit;
At least one described second delay cell, for carrying out prolonging for unit interval to the signal from the state conversion unit Late, and by the signal after the delay by least one second delay cell it is input to the state conversion unit;
The state conversion unit, for using loop cycle matrix to from first delay cell and it is described at least one The signal of second delay cell is weighted processing, and the signal after weighting is corresponded to input respectively to the adder and described At least one second delay cell;The parameter of the loop cycle matrix is rung according to the frequency domain impulse of the simulation low-pass filter The coefficient of function is answered to obtain, the coefficient of the frequency domain impulse response function meets following condition:According to the frequency domain impulse response Any two row for the calculation matrix that function and sampling function are obtained are uncorrelated;
One of them second delay cell in first delay cell, or at least one described second delay cell, is also used Signal after by delay is input to the sampler being connected with the simulation low-pass filter.
2. simulation low-pass filter according to claim 1, it is characterised in that in the frequency domain impulse response function H at j momentj (z) it is:
Wherein, i be the moment numbering, T be the loop cycle matrix cycle, N for A (t) row vector or column vector number, λ′k,i,jFor Ψ (T, 0) k-th of characteristic value, λ "k,i,jFor Ψ (i-1,0) Ψ (T, k-th of characteristic value j), λ " 'k,i,jFor Ψ (i-1, k-th of characteristic value j);
Wherein A (t) is loop cycle matrix.
3. simulation low-pass filter according to claim 2, it is characterised in that any two row of calculation matrix are uncorrelated Specifically represented with equation below:
Wherein, F Φi=D (z) Hi(z), Hi(z) it is the frequency domain impulse response function at the i moment, D (z) is sampling function;ΦiFor D (z)Hi(z) function after Fourier transformation, that is, the corresponding calculation matrix of moment i, F refer to Fourier transformation;FΦj =D (z) Hj(z), ΦjFor D (z) Hj(z) function after Fourier transformation, that is, the corresponding calculation matrix of moment j, F is Refer to Fourier transformation;(FΦi)HFor to F ΦiConjugate transposition is sought, | | F Φi||2To seek F ΦiMould 2 norms, | | F Φj||2For Seek F ΦjMould 2 norms, the δ be waviness tolerance value, the VlFor calculation matrix F ΦjIt is any two row.
4. simulation low-pass filter according to claim 2, it is characterised in that the condition also includes:The limit of system In unit circle.
5. simulation low-pass filter according to claim 4, it is characterised in that the limit of the system is used in unit circle Equation below is represented:
Wherein, | | λ 'k,i,j||2To seek λ 'k,i,jMould 2 norms.
6. simulation low-pass filter according to claim 2, it is characterised in that the condition also includes:System pole is big In or equal to preset value.
7. simulation low-pass filter according to claim 6, it is characterised in that the system pole is more than or equal to default Value is represented by equation below:
min{||λ′k,i,j||2}≥γ
Wherein, the γ is preset value, represents lower bound of the limit apart from origin.
8. simulation low-pass filter according to claim 1, it is characterised in that the condition also includes:The cycle is followed The parameter of ring matrix is within a preset range.
9. a kind of analog information converter, it is characterised in that the analog information converter include frequency mixer, sampler and The simulation low-pass filter of claim 1 to 8 any one, the frequency mixer is connected with the simulation low-pass filter, described Simulation low-pass filter is connected with the sampler;
The frequency mixer, for initial signal to be mixed with random sequence signal, obtains mixed frequency signal;
The simulation low-pass filter, for being filtered to the mixed frequency signal, obtains filtering signal;
The sampler, for being sampled to the filtering signal.
10. a kind of simulation low-pass filter, it is characterised in that including:
Processor;
Memory;And COM1;
Wherein, the processor is used for the instruction for performing the memory storage, to perform following steps in execute instruction:
Received by the COM1 come the mixed frequency signal of self-mixing device, the mixed frequency signal is initial signal and random sequence Signal after signal mixing, the frequency mixer is connected with simulation low-pass filter;
The mixed frequency signal is overlapped with the signal after weighting and obtains composite signal;
The delay of unit interval is carried out to the composite signal, and the delay of unit interval is carried out to the signal after the weighting;
The signal after processing obtains the weighting is weighted to the signal by delay using loop cycle matrix;It is described The parameter of loop cycle matrix is obtained according to the coefficient of the frequency domain impulse response function of the simulation low-pass filter, the frequency domain The coefficient of impulse response function meets following condition:The measurement square obtained according to the frequency domain impulse response function and sampling function Any two row of battle array are uncorrelated;
Signal after delay is input to by the sampler being connected with the simulation low-pass filter by the COM1.
11. simulation low-pass filter according to claim 10, it is characterised in that in the frequency domain impulse response letter at j moment Number Hj(z) it is:
Wherein, i be the moment numbering, T be the loop cycle matrix cycle, N for A (t) row vector or column vector number, λ′k,i,jFor Ψ (T, 0) k-th of characteristic value, λ "k,i,jFor Ψ (i-1,0) Ψ (T, k-th of characteristic value j), λ " 'k,i,jFor Ψ (i-1, k-th of characteristic value j);
Wherein A (t) is loop cycle matrix.
12. simulation low-pass filter according to claim 11, it is characterised in that the calculation matrix any two arranges not phase Pass is specifically represented with equation below:
Wherein, F Φi=D (z) Hi(z), Hi(z) it is the frequency domain impulse response function at the i moment, D (z) is sampling function;ΦiFor D (z)Hi(z) function after Fourier transformation, that is, the corresponding calculation matrix of moment i, F refer to Fourier transformation;FΦj =D (z) Hj(z), ΦjFor D (z) Hj(z) function after Fourier transformation, that is, the corresponding calculation matrix of moment j, F is Refer to Fourier transformation;(FΦi)HFor to F ΦiConjugate transposition is sought, | | F Φi||2To seek F ΦiMould 2 norms, | | F Φj||2For Seek F ΦjMould 2 norms, the δ be waviness tolerance value, the VlFor calculation matrix F ΦjIt is any two row.
13. simulation low-pass filter according to claim 10, it is characterised in that the condition also includes:The pole of system Point is in unit circle.
14. simulation low-pass filter according to claim 10, it is characterised in that the condition also includes:System pole More than or equal to preset value.
15. simulation low-pass filter according to claim 10, it is characterised in that the condition also includes:The cycle The parameter of circular matrix is within a preset range.
16. a kind of filtering method, it is characterised in that methods described includes:
Simulation low-pass filter receives the mixed frequency signal for carrying out self-mixing device, and the mixed frequency signal is that initial signal is believed with random sequence Signal after number mixing, the frequency mixer is connected with simulation low-pass filter;
The mixed frequency signal is overlapped and obtains composite signal by the simulation low-pass filter with the signal after weighting;
The simulation low-pass filter carries out the delay of unit interval to the composite signal, and the signal after the weighting is entered The delay of row unit interval;
The simulation low-pass filter is weighted processing to the signal by delay using loop cycle matrix and obtains institute State the signal after weighting;The parameter of the loop cycle matrix is according to the frequency domain impulse response function of the simulation low-pass filter Coefficient obtain, the coefficient of the frequency domain impulse response function meets following condition:According to the frequency domain impulse response function and Any two row for the calculation matrix that sampling function is obtained are uncorrelated;
Signal after delay is input to the sampler being connected with the simulation low-pass filter by the simulation low-pass filter.
17. filtering method according to claim 16, it is characterised in that in the frequency domain impulse response function H at j momentj(z) For:
Wherein, i be the moment numbering, T be the loop cycle matrix cycle, N for A (t) row vector or column vector number, λ′k,i,jFor Ψ (T, 0) k-th of characteristic value, λ "k,i,jFor Ψ (i-1,0) Ψ (T, k-th of characteristic value j), λ " 'k,i,jFor Ψ (i-1, k-th of characteristic value j);
Wherein A (t) is loop cycle matrix.
18. filtering method according to claim 17, it is characterised in that any two row of calculation matrix are uncorrelated specific Represented with equation below:
Wherein, F Φi=D (z) Hi(z), Hi(z) it is the frequency domain impulse response function at the i moment, D (z) is sampling function;ΦiFor D (z)Hi(z) function after Fourier transformation, that is, the corresponding calculation matrix of moment i, F refer to Fourier transformation;FΦj =D (z) Hj(z), ΦjFor D (z) Hj(z) function after Fourier transformation, that is, the corresponding calculation matrix of moment j, F is Refer to Fourier transformation;(FΦi)HFor to F ΦiConjugate transposition is sought, | | F Φi||2To seek F ΦiMould 2 norms, | | F Φj||2For Seek F ΦjMould 2 norms, the δ be waviness tolerance value, the VlFor calculation matrix F ΦjIt is any two row.
19. filtering method according to claim 16, it is characterised in that the condition also includes:The limit of system is in list In circle of position.
20. filtering method according to claim 16, it is characterised in that the condition also includes:System pole be more than or Equal to preset value.
21. filtering method according to claim 16, it is characterised in that the condition also includes:The loop cycle square The parameter of battle array is within a preset range.
CN201610067654.5A 2016-01-29 2016-01-29 Analog low-pass filter, analog information converter and filtering method Active CN107026650B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610067654.5A CN107026650B (en) 2016-01-29 2016-01-29 Analog low-pass filter, analog information converter and filtering method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610067654.5A CN107026650B (en) 2016-01-29 2016-01-29 Analog low-pass filter, analog information converter and filtering method

Publications (2)

Publication Number Publication Date
CN107026650A true CN107026650A (en) 2017-08-08
CN107026650B CN107026650B (en) 2020-10-09

Family

ID=59525175

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610067654.5A Active CN107026650B (en) 2016-01-29 2016-01-29 Analog low-pass filter, analog information converter and filtering method

Country Status (1)

Country Link
CN (1) CN107026650B (en)

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1813418A (en) * 2003-05-23 2006-08-02 模拟设备股份有限公司 Sparse echo canceller
US20080169948A1 (en) * 2004-12-03 2008-07-17 Kouji Okamoto Filter Adjustment Circuit
CN102549928A (en) * 2009-09-30 2012-07-04 熵敏通讯股份有限公司 Method and apparatus for bandpass digital to analog converter
CN102571034A (en) * 2011-12-30 2012-07-11 北京邮电大学 Random circulant matrix-based analog compressed sensing sampling method and system
CN102723952A (en) * 2012-06-27 2012-10-10 华为技术有限公司 Analog/digital conversion data transmission method, transmitting device and analog/digital conversion chip
CN102931998A (en) * 2012-10-11 2013-02-13 上海交通大学 Analog-information sampling method in cloud manufacturing platform
CN103178853A (en) * 2013-03-21 2013-06-26 哈尔滨工业大学 Compressive-sensing-based sparse signal under-sampling method and implementation device
CN104218954A (en) * 2014-08-28 2014-12-17 中国电子科技集团公司第二十九研究所 Method and device for compressed sampling of broadband array antenna
CN104467858A (en) * 2013-09-25 2015-03-25 中国科学院深圳先进技术研究院 Time domain integration sampling method and sampling circuit
US9197283B1 (en) * 2014-12-18 2015-11-24 Raytheon Company Reconfigurable wideband channelized receiver

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1813418A (en) * 2003-05-23 2006-08-02 模拟设备股份有限公司 Sparse echo canceller
US20080169948A1 (en) * 2004-12-03 2008-07-17 Kouji Okamoto Filter Adjustment Circuit
CN102549928A (en) * 2009-09-30 2012-07-04 熵敏通讯股份有限公司 Method and apparatus for bandpass digital to analog converter
CN102571034A (en) * 2011-12-30 2012-07-11 北京邮电大学 Random circulant matrix-based analog compressed sensing sampling method and system
CN102723952A (en) * 2012-06-27 2012-10-10 华为技术有限公司 Analog/digital conversion data transmission method, transmitting device and analog/digital conversion chip
CN102931998A (en) * 2012-10-11 2013-02-13 上海交通大学 Analog-information sampling method in cloud manufacturing platform
CN103178853A (en) * 2013-03-21 2013-06-26 哈尔滨工业大学 Compressive-sensing-based sparse signal under-sampling method and implementation device
CN104467858A (en) * 2013-09-25 2015-03-25 中国科学院深圳先进技术研究院 Time domain integration sampling method and sampling circuit
CN104218954A (en) * 2014-08-28 2014-12-17 中国电子科技集团公司第二十九研究所 Method and device for compressed sampling of broadband array antenna
US9197283B1 (en) * 2014-12-18 2015-11-24 Raytheon Company Reconfigurable wideband channelized receiver

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
MICHAEL TRAKIMAS 等: "A Compressed sensing analog-to-information converter with edge-triggered SAR ADC Core", 《2012 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS (ISCAS)》 *
RAY MALEH 等: "Analog-to-Information and the Nyquist Folding Receiver", 《IEEE JOURNAL ON EMERGING AND SELECTED TOPICS IN CIRCUITS AND SYSTEMS》 *
周权 等: "基于压缩感知的模拟信息转换器仿真", 《物联网技术》 *

Also Published As

Publication number Publication date
CN107026650B (en) 2020-10-09

Similar Documents

Publication Publication Date Title
Wyglinski et al. Software-defined radio for engineers
Asmussen et al. Fitting phase-type distributions via the EM algorithm
CN109101272A (en) Processing with Neural Network device and its method for executing matrix multiple instruction
CN112073221B (en) Method and device for realizing network node sequencing
CN108205519A (en) The multiply-add arithmetic unit of matrix and method
CN106456115A (en) Ultrasonic imaging compression methods and apparatus
CN107634768B (en) FPGA-based (field programmable Gate array-based) MWC (Multi-media wall) compression sampling broadband digital receiver PDW (Polymer dispersed W) forming method
CN108960418A (en) Processing with Neural Network device and its method for executing matrix-vector multiplication instruction
Riznyk Multi-modular Optimum Coding Systems Based on Remarkable Geometric Properties of Space
CN104424507B (en) Prediction method and prediction device of echo state network
CN106093908A (en) A kind of radar target detection method based on piecemeal segmentation AIC model
JPH05189471A (en) Butterfly-shaped operator
CN111310589A (en) Fault diagnosis method and device for mechanical system and terminal
Liu et al. Wideband spectrum sensing based on single-channel sub-Nyquist sampling for cognitive radio
CN107026650A (en) A kind of simulation low-pass filter, analog information converter and filtering method
CN112131638B (en) Wind-induced dynamic characteristic type determination method of large-span roof structure and terminal equipment
Karube On the local cross-sections in locally compact groups
Quirk et al. A wide-band high-resolution spectrum analyzer
JPWO2006106713A1 (en) A signal approximation system that minimizes the worst-case error measure.
Sudakshina Analog and Digital Communications
CN107147396A (en) Signal sampling method, signal sampling system and signal sampling device
Asl et al. Multichannel sampling of multidimensional parametric signals
KR102505022B1 (en) Fully parallel fast fourier transform device
Baghaie et al. Computing Haar transform using algebraic integers
Kumar et al. Tour Planning Mechanism in FFT

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant