CN102095933A - Phase test device and control method thereof - Google Patents

Abstract

The invention discloses a phase test device and a control method thereof, and relates to the technical field of digital signal processing. The phase test device comprises a conditioning circuit, an A/D converter, a digital signal processor, an output driver and a display circuit of the output driver, wherein an analog signal is accessed to the conditioning circuit, the conditioning circuit adjusts the amplitude of the analog signal according to a first threshold to obtain a first analog signal, the first analog signal is sampled by the A/D converter to obtain a sample sequence, the sample sequence is accessed into the digital signal processor through a parallel digital input manner, an estimated phase value of the sample sequence is obtained through processing of the digital signal processor, and the estimated phase value of the sample sequence is output and displayed by the output driver and the display circuit of the output driver. The phase test device can realize accurate estimation of instantaneous phase of the center sample point by using shorter sample data; the phase test device can realize phase estimation without frequency estimation; without frequency estimation, error of frequency estimation is avoided, the cost of hardware and the requirement on complexity are reduced, and the requirements in actual application are met.

Description

A kind of survey is installed and control method mutually

Technical field

The present invention relates to digital signal processing technique field, particularly a kind of survey is installed and control method mutually.

Background technology

In engineering fields such as communication, instrument, electric power, optical application and fault diagnosises, exist in a large number the phase place of signal is carried out high precision, estimation problem efficiently and fast.For example in adopting QAM and PSK communication system for modulation, only the carrier wave initial phase in each code-element period is carried out Fast estimation, could correctly recover planisphere, realize being correctly decoded at receiving end ^[1]In all kinds of instrument, phasometer is common a kind of, and the basic function of phasometer is exactly accurately to calculate each phase place constantly of signal ^[2]And, require accurately to measure the phase differential between power frequency component in two network systems in the combined floodgate occasion that is incorporated into the power networks of electric system; In addition in harmonic analysis in power system; be subjected to the influence of generator windings technology, various non-linear power equipment load factors such as (as various thyristor rectifier devices, converter plants); mains frequency can be offset usually, measures phase place, frequency and the amplitude of 50Hz first-harmonic and each harmonic when this will be realistic ^[3]Phase type high-precision laser range-finding instrument is exactly that continuous amplitude-modulated signal is past on testing distance by measuring, the phase delay that is produced is broadcast in back pass, measures distance indirectly ^[4]In the fault diagnosis of rotating machinery, usually need to come the failure judgement classification by vibration signal being carried out analysis of spectrum (comprising spectral amplitude and phase spectrum), yet sometimes spectral amplitude often corresponding two or three explanations (from the signal spectrum that rotating machinery collects, exist 2 bigger frequency multiplication components such as certain, showing then that this machine may exist misaligns or bending shaft or the loosening problem of machinery, be any situation on earth, clear hardly), if at this moment in conjunction with phase analysis, just be easy to make accurately and judge ^[5]In a word, the estimation problem of signal phase is prevalent in the production link that is closely related with national economy, thereby develops high performance phase estimation method and have very high construction value and economic worth.

Phase measurement and estimation approach are a lot, and tradition relies on the method for analog device, as: vector method, diode phase demodulation method, pulse counting method etc., its measuring system complexity needs dedicated devices, the hardware cost height.In recent years, phase estimation develops to the digitizing direction gradually, and its advantage is that hardware cost is low, adaptability is strong, only needs general-purpose devices such as single-chip microcomputer, DSP and FPGA just can finish, only need reprogramming to get final product to different measuring objects, and its precision generally is higher than analog measurement.Therefore, the accurate phase estimation method of a selected cover is crucial.

Yet for developing high-precision phase measurement method (being method for parameter estimation), it is far from being enough only going to consider a problem in the engineering field.Because no matter be in which kind of application scenario, no matter be to measure which kind of physical quantity also, the form of expression of signal all is the discrete observation data after sampling, if want accurately, fast, from observation data, extract parameter information effectively, this just relates to a lot of theoretical questions, the knowledge that is contained can cover digital signal processing, input and estimation theory, information theory, Probability Theory and Math Statistics, a plurality of ambits such as stochastic process, has only the understanding of intensification to the rudimentary knowledge in these fields, propose the measure of some parameters optimization estimated performances from new angle, just might develop the more good method of performance.And existing digitizing phase estimation mainly comprises following several method:

(1) pulse counting method

This is the most frequently used phase measurement, it surveys facies principle as shown in Figure 1, main process by: produce with the reference sine wave of survey sinusoidal signal same frequency f and known first phase, respectively this two-way sine wave is carried out step-by-step counting then and (supposes that its basic count cycle is T ₀), calculate its time-delay nT according to its step-by-step counting difference (supposing n pulse difference) ₀, and then calculate its phase differential 2 π fnT ₀, add this phase differential by fixed phase, can obtain phase estimation.Can find out that the defective of this method is: (a) need to produce reference sine wave, and require reference signal in full accord,, then can bring very serious deviation to measurement result if be slightly offset with the frequency of the sinusoidal signal of surveying; (b) need special generation pulse counting signal, and measuring accuracy depends on basic pulse width T to a great extent ₀, only reduce T ₀Could improve measuring accuracy, this is very high to requirements on hardware equipment.

(2) Hilbert transform method ^[6]

Signal x (t)=cos (2 π ft+ θ for the cosine class ₀), the phase theta during for estimation t=0 ₀, with sample frequency f _sObtain sequence x (n)=cos (2 π f/f after x (t) sampled _sN+ θ ₀), n=0,1 ..., N-1; If being done Hilbert transform, x (n) can get its analytic signal , n=0,1 ..., N-1 is then according to the available θ of following formula ₀Estimator

${\widehat{\mathrm{\θ}}}_0=\arctan \frac{\hat{x}\left(0\right)}{x\left(0\right)}---\left(1\right)$

It is very high that yet the Hilbert transform method requires sample frequency, as signal frequency f=kf _s/ N, k ∈ z ⁺The time, the phase place that following formula records is accurately, in case sample frequency f _sDepart from slightly, make this integral multiple relation be false, the estimation of formula (1) is just incorrect; The noiseproof feature of the method is poor.

(3) sinusoidal curve fitting process ^[7～9]

John Kuffel has proposed the sinusoidal curve fitting process ^[7,8], this method can obtain very high estimated accuracy.It is divided into four parameters (frequency, amplitude, phase place and DC component all unknown) method and three parametric methods (except that frequency is known, other 3 all the unknowns) two kinds of situations, document [9] is pointed out, four parameter sinusoidal curve fit procedure are not closed linear process, still do not have definite mathematical formulae and can directly calculate fitting parameter, if it is improper that the match starting condition is selected, easily make iterative process disperse or converge to local optimum, and match needs a large amount of operation times.For raising the efficiency, document [9] proposes first estimation frequency, carries out the sinusoidal curve match (for closed linear process, and absolute convergence) of three parameters again.Yet the method requires Frequency Estimation very accurate, otherwise a direct result can be brought into estimated frequency error in the phase estimation exactly, and in addition, the method still can only be carried out phase estimation to simple signal.

(4) correlation method ^[10～11]

Document [10-11] has adopted correlation method to survey phase, and the hypothetical reference sample sequence is x (n)=cos (2 π f/f _sN), n=0,1 ..., N-1, sequence to be measured is y (n)=cos (2 π f/f _sN+ θ ₀), n=0,1 ..., N-1; For estimating the first phase that y (n), need obtain the autocorrelation function R of x (n) and y (n) respectively _Xx(τ), R _Yy(τ), and both cross correlation function R _Xy(τ), can obtain θ according to formula (2) ₀Estimation

${\widehat{\mathrm{\θ}}}_0=\arccos \frac{R_{\mathrm{xy}}\left(0\right)}{\sqrt{R_{\mathrm{xx}}\left(0\right)R_{\mathrm{yy}}\left(0\right)}}---\left(2\right)$

Yet can find out that from formula (2) correlation method is surveyed mutually still need one tunnel and the duplicate reference signal of measured signal frequency, and still only be suitable for the phase estimation of simple signal; In addition, experiment showed, that the noiseproof feature of the method is very poor.

(5) all kinds of phase estimation methods of proofreading and correct based on the frequency spectrum of discrete Fourier transform (DFT)

Based on discrete Fourier transform (DFT) (Discrete Fourier Transform, phase estimation method DFT) be in recent years engineering circle and academia studies the most hotly, also be the widest parameter estimation method of practical application.Reason is, when one, DFT exponent number are 2 power power, can be realized by its fast algorithm FFT (Fast Fourier Transform, fast fourier transform), has the incomparable convenience of other algorithm, succinct calculating advantage; Its two, in theory, when N is enough big, can obtain the close approximation that maximum likelihood is separated if the discrete sampling value is carried out DFT; They are three years old, various abundant information can directly obtain reflection in the result of DFT, distribution as the DFT position of spectral line has reflected the signal frequency size, and DFT result is a plural number, thereby implied (this are that other orthogonal transformation is not available, as discrete cosine transform, Walsh transform, Karhunent-Loeve transformation (Karhunen-Loeve transformation), Wigner-ville conversion etc.) such as phase informations of signal.Though the fence effect of DFT has limited its frequency resolution, but its fence effect can be solved by the method for the DFT spectral line being carried out interpolation, so just can derive multiple discrete spectrum correction method, document [12] has been summed up 4 class spectrum correction methods, comprises the energy barycenter method ^[13], ratioing technigue ^[14-15], DFT+FFT spectrum refinement method ^[16], phase difference method ^[17-19], these methods all can be carried out more accurate phase estimation to the signal that has multiple radio-frequency component, thereby are widely used in fields such as vibration analysis and fault detect, harmonic analysis in power system, dielectric loss angular measurement, radar velocity measurement.

Yet can prove with theory by experiment, there is following defective in all kinds of phase estimation methods of proofreading and correct based on the DFT frequency spectrum: these DFT spectrum correction methods are when estimating phase place, all be by first estimated frequency, obtain carrying out again after the frequency departure value that the step of phase estimation carries out, thereby can directly be brought into the error of Frequency Estimation in the phase estimation like this.

(6) whole phase FFT is surveyed the phase method

Whole phase FFT is a kind of novel spectral analysis method that patent applicant Wang Zhao China proposes at patent documentation [20], document [21] is pointed out, one time whole phase FFT combines N the son spectrum mechanism of compensation mutually, thereby whole phase FFT has good inhibition spectrum leakage characteristics and " phase invariant ", when estimating phase place, whole phase FFT need not to carry out Frequency Estimation can direct estimation go out the signal phase value.Yet document [21,22] points out that the whole phase FFT on N rank need expend 2N-1 input sample, and the instantaneous phase that just can obtain the center sampling point is estimated.Though thereby whole phase FFT precision height, it is many slightly to expend sampling point, inapplicable for the phase measurement occasion of short sample low frequency signal.

List of references

[1] English process plate [M] .Publishing of John G.Proakis.Digital Communications:Fourth Edition house of electronics industry, Beijing, 2006,9:336-340

[2]Marcin，M.R.，Digital?receiver?phase?meter?for?LISA[J].IEEE?Transactions?onInstrumentation?and?Measurement，2005，54(6)：2446-2453

[3]Zhang?Fusheng，Geng?Zhongxing，Yuan?Wei.The?algorithm?of?interpolatingwindowed?FFT?for?harmonic?analysis?of?electric?power?system[J].IEEETransactions?on?Power?Delivery，2001，16(2)：160-164.

[4]Bykov?M.M.，Tyurin?V.S.，Tyurin?S.V..Signals?Modeling?in?Laser?DistanceMeasurements[C].8-th?International?Conference?on?Laser?and?Fiber-OpticalNetworks?Modeling，June?2006：297-300

[5] Zuo Jinggang. the phase analysis diagnosis [J] of equipment failure. Chinese equipment control, 5 phases of calendar year 2001: 37-39

[6] Huang Dishan, Lu Naiyan, Tong Zhongfang, Cheng Yaodong [J]. use Hilbert transform and extract phase signal. vibration and dynamic test, 1988:68-71

[7]John?Kuffel，Terence?R?McComb，Rysard?Malewski.Compa-rative?methods?forcalculating?the?best-fit?sinusoid?to?the?digital?record?of?a?high-purity?sine?wave[J].IEEE?Trans.on?instrumentation?and?measurement，1987，36(3)：418-422

[8]Terence?R?McComb，John?Kuffel?and?Bryant?C?Le?Roux.A?comparative?evaluationof?some?practical?algorithms?used?in?the?effective?bits?test?of?waveformrecorder[J].IEEE?Trans.on?Instrumentation?and?measurement，1989，38(1)：37-42

[9] Liang Zhiguo magnifies and controls, grandson's Jing space, Li Xinliang. the fast algorithm [J] of four parameter sine wave curve matches. and instrumentation technology, 2006,26 (1): 4-7

[10] Li Guochao, Huang Qinghua. based on the phase difference measurement of relative theory and the design of virtual phasometer [J]. microcomputer information, 2007,23 (1): 185-186

[11] Wu Junqing. the digitized measurement research [J] of phase differential. application foundation and engineering science journal, 2005,13 (1): 99-104

[12] Zhu Xiaoyong, the fourth health. the comprehensive comparison [J] of discrete spectrum correction method. signal Processing, 2001,17 (1): 91-97

[13] fourth health, Jiang Liqi. the energy barycenter correction method [J] of discrete spectrum. vibration engineering journal, 2001,14 (3): 354-359

[14]Xie?M，Ding?K.Correction?for?the?frequency，amplitude?and?phase?in?FFT?ofharmonic?signal[J].Mechanical?System?and?Signal?Processing，1996，10(2)：211-221

[15] Liu Min, Wang Keying. based on the high precision FFT frequency analysis [J] of the bimodal spectral line interpolation of windowing. electrical measurement and instrument, 2006,43 (3): 112-116

[16] Liu Jinming should cherish firewood .FFT and compose the Fourier transform method [J] that continuous refinement is analyzed. vibration engineering journal, 1995,18 (2): 162-166

[17] fourth health, Luojiang County is triumphant, Xie Ming. discrete spectrum time shift phase difference correction method [J]. and applied mathematics and mechanics, 2002,23 (7): 729-735

[18] fourth health, Zhong Shuncong. general discrete spectrum phase difference correction method [J]. electronic letters, vol, 2003,31 (1): 142-145

[19] Huang Zhiyun, Xu Kejun. based on the research [J] of the spectrum correcting method of phase differential. vibration with impact 2005,24 (2): 77-79

[20] Wang Zhaohua, Hou Zhengxin. whole phase FFT arrangements for analyzing frequency [P]. utility model patent, application (patent) number: 200420028959.8

[21] Wang Zhaohua, Huang Xiangdong, Yang Wei. whole phase FFT phase measurement [J]. world technology research and development, in August, 2007, the 29th volume the 4th phase: 28-32.

[22] Wang Zhaohua, Huang Xiangdong. based on the phase measurement principle and the application [J] thereof of full phase spectral analysis. data acquisition and processing, 2009,24 (6): 777-782

Summary of the invention

In order to adopt less sampling point to realize the scope that accurate estimation, reduction computation complexity, expansion to center sampling point instantaneous phase are used, the invention provides a kind of survey and install mutually and control method,

A kind of survey is installed mutually, and described device comprises: modulate circuit, A/D converter, digital signal processor and output driving and display circuit thereof,

Simulating signal inserts described modulate circuit, described modulate circuit is adjusted the amplitude of described simulating signal according to first threshold, obtain first simulating signal, described first simulating signal obtains sample sequence through described A/D converter sampling, described sample sequence inserts described digital signal processor with the form of Parallel Digital input, obtain the phase estimation value of sample sequence through the processing of described digital signal processor, the phase estimation value of described sample sequence is by described output driving and display circuit output thereof and show.

A kind of survey phase control method said method comprising the steps of:

(1) sets sample frequency f _sWith analysis of spectrum exponent number N;

(2) with described sampling frequency f _sInput simulating signal x (t) is carried out equal interval sampling, obtain

Be total to N+1 discrete sample;

(3) a described N+1 discrete sample is divided into two sections, first section discrete sample is by the top n discrete sample

Form, second section discrete sample is by N the discrete sample in back

Form;

(4) with described first section discrete sample ring shift left

, obtain first section discrete sample after moving to left; With described second section discrete sample ring shift left

Obtain second section discrete sample after moving to left;

(5) first section discrete sample after described the moving to left done the DFT conversion, second section discrete sample after described the moving to left done the DFT conversion,, obtain spectrum Y (k) the summation of the result after the secondary DFT conversion, wherein, Y (k), k=0 ..., N-1};

(6), obtain the amplitude peak of spectrum Y (k) to described spectrum Y (k) delivery value;

(7) with the phase value ang[Y (k*) of the amplitude peak correspondence of described spectrum Y (k)] as the phase estimation value of center sampling point x (0);

(8) judge whether described phase estimation value equals second threshold value, if output drives and display circuit is exported described phase estimation value; If not, adjust described analysis of spectrum exponent number N, re-execute step (2).

In step (1) afterwards, step (2) before, described method also comprises: eliminate the flip-flop among the described simulating signal x (t).

The control method that a kind of survey is installed mutually said method comprising the steps of:

(9) set sample frequency f _sWith analysis of spectrum exponent number N;

(10) with described sample frequency f _sInput simulating signal x (t) is carried out equal interval sampling, obtain

Be total to N+1 discrete sample;

(11) with the sample in the described N+1 discrete sample

Take advantage of 2, first sampling point x (N/2) with last sampling point x (N/2) addition, is obtained N sampling point y (0)～y (N-1);

$y\left(n\right)=\left\{\begin{array}{c}2x\left(n\right),n\&Element;[0,\frac{N}{2}-1]\\ x(-\frac{N}{2})+x\left(\frac{N}{2}\right),n=\frac{\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="HDA0000036676480000042.png,HDA0000036676480000051.png"\; state="\{\{state\}\}">N</figure-callout>}}{2}\\ 2x(n-N),n\&Element;[\frac{\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="HDA0000036676480000042.png,HDA0000036676480000051.png"\; state="\{\{state\}\}">N</figure-callout>}}{2}+1,N-1]\end{array}\right.$

(12) described N sampling point y (0)～y (N-1) is done DFT 1 time, obtain spectrum Y (k);

(13), obtain the amplitude peak of spectrum Y (k) to described spectrum Y (k) delivery value;

(14) with the phase value ang[Y (k*) of the amplitude peak correspondence of described spectrum Y (k)] as the phase estimation value of center sampling point x (0);

(15) judge whether described phase estimation value equals second threshold value, if output drives and display circuit is exported described phase estimation value; If not, adjust described analysis of spectrum exponent number N, re-execute step (10).

In step (9) afterwards, step (10) before, described method also comprises: eliminate the flip-flop among the simulating signal x (t).

The beneficial effect of technical scheme provided by the invention is:

The invention provides a kind of survey and install mutually and control method, the present invention can realize the accurate estimation to center sampling point instantaneous phase with short sample data; Do not need estimated frequency just can realize phase estimation; Owing to do not carry out Frequency Estimation, avoided the error of Frequency Estimation, make that the precision of phase estimation value is very high; After adopting further improvement, only carry out DFT one time, reduced calculated amount, computation complexity is low; Owing to introduced DFT, make that most of noise energy is shared on the very wide non-signal band in the frequency domain, greatly reduce the interference of noise to phase measurement, higher noise resisting ability is arranged; Need not reference signal and can measure phase place, reduced, satisfied the needs in the practical application the cost of hardware and the requirement of complexity.

Description of drawings

Fig. 1 surveys facies principle figure for the pulse counting method that prior art provides;

The structural representation that Fig. 2 installs mutually for survey provided by the invention;

Fig. 3 is the measurement structure synoptic diagram based on quarter-phase DFT provided by the invention;

Fig. 4 is the process flow diagram of survey phase control method provided by the invention;

Fig. 5 is the measurement structure synoptic diagram of the equivalence of quarter-phase DFT provided by the invention;

Fig. 6 is another process flow diagram of survey phase control method provided by the invention;

Fig. 7 is two son spectrums provided by the invention and quarter-phase DFT spectrum contrast figure;

Fig. 8 surveys phase standard deviation trial curve figure under the noise provided by the invention.

Embodiment

For making the purpose, technical solutions and advantages of the present invention clearer, embodiment of the present invention is described further in detail below in conjunction with accompanying drawing.

In order to adopt less sampling point to realize the scope that accurate estimation, reduction computation complexity, expansion to center sampling point instantaneous phase are used, the embodiment of the invention provides a kind of survey to install mutually, referring to Fig. 2, sees for details hereinafter and describes:

This survey device mutually comprises: modulate circuit, A/D converter, digital signal processor and output driving and display module thereof,

Simulating signal inserts modulate circuit, modulate circuit is adjusted the amplitude of simulating signal according to first threshold, obtain first simulating signal, first simulating signal obtains sample sequence through the A/D converter sampling, sample sequence is with the form incoming digital signal processor of Parallel Digital input, obtain the phase estimation value of sample sequence through the processing of digital signal processor, the phase estimation value of sample sequence is driven by output and display module output and showing.

Wherein, modulate circuit is adjusted the amplitude of simulating signal according to first threshold, make the changes in amplitude of simulating signal meet the requirement of the quantizing range of A/D converter, promptly, first threshold is set according to the quantizing range of A/D converter in the practical application, during specific implementation, the embodiment of the invention does not limit this.

A kind of survey phase control method, referring to Fig. 3 and Fig. 4, this method may further comprise the steps:

101: set sample frequency f _sWith analysis of spectrum exponent number N;

Wherein, can be according to requirement of actual application, for example: the concrete measurement requirement of medical science and military affairs etc., that carries out sample frequency and analysis of spectrum exponent number N determines that sample frequency has determined the output clock of digital signal processor; During practical engineering application, preferably selecting N is 2 power power, and DFT can use FFT (Fast FourierTransform, fast fourier transform) to substitute like this, and spent complex multiplication calculation times is from N ²The inferior N/2log that reduces to ₂N time, calculated amount reduces greatly.

102: with sample frequency f _sInput simulating signal x (t) is carried out equal interval sampling, obtain Be total to N+1 discrete sample;

This N+1 discrete sample promptly

103: N+1 discrete sample is divided into two sections, and first section discrete sample is by the top n discrete sample Form, second section discrete sample is by N the discrete sample in back Form;

In order to measure the phase place of center sampling point x (0), this N+1 discrete sample is divided into two sections, first section discrete sample is by the top n discrete sample

Form, second section discrete sample is by N the discrete sample in back Form.

104: with first section discrete sample ring shift left

, obtain first section discrete sample after moving to left; With second section discrete sample ring shift left

Obtain second section discrete sample after moving to left;

With first section discrete sample ring shift left

The position promptly becomes

With second section discrete sample ring shift left

The position promptly becomes

105: first section discrete sample after moving to left done the DFT conversion, second section discrete sample after moving to left done the DFT conversion,, obtain spectrum Y (k) to the summation of the result after the DFT conversion;

Spectrum Y (k) is obtained in summation to the result after the DFT conversion, and Y (k), k=0 ..., N-1}.

106:, obtain the amplitude peak of spectrum Y (k) to spectrum Y (k) delivery value;

107: the phase value ang[Y (k*) that will compose the amplitude peak correspondence of Y (k)] as the phase estimation value of center sampling point x (0);

108: judge whether phase estimation value equals first threshold, if output drives and display module is exported the phase estimation value of sample sequence; If not, adjust analysis of spectrum exponent number N, re-execute step 102.

Wherein, first threshold is set according to the engineering demand in the practical application, and during specific implementation, the embodiment of the invention does not limit this.

In the practical application, if there is flip-flop in simulating signal, can reduce the precision of measurement, in order further to improve precision, after step 101, before the step 102, this method also comprises: eliminate the flip-flop in the simulating signal for this reason.Wherein, by calculating the mean value of sampling point, just can obtain the flip-flop in the simulating signal.

Owing to be split as at Fig. 3 two different reference position elements arranged in two overlapping segmentations (be respectively x (N/2) and x (N/2+1)), in digital signal processing, what in fact the difference of reference position reflected is the difference of signal initial phase, so with the analysis of spectrum in the empty frame of Fig. 3 be referred to as quarter-phase DFT (double-phaseDiscrete Fourier Transform, dpDFT).

With N=8 is example, and input signal is x (n)=cos (ω ₀N+ θ ₀), n ∈ [4,4], wherein ω ₀=2.3 Δ ω, Δ ω=2 π/8, θ ₀=40 °, input sample sequence x (4)～x (4) is [0.9703 0.0087-0.97440.4462 0.7660-0.8039-0.3907 0.9863-0.0698]

Above-mentioned sampling point is divided into two sections overlapping sample x ₁', x ₂' be

x ₁′＝[0.9703 0.0087 -0.9744 0.4462 0.7660 -0.8039 -0.39070.9863]

x ₂′＝[0.0087 -0.9744 0.4462 0.7660 -0.8039 -0.3907 0.9863-0.0698]

With first section discrete sample x ₁' ring shift left

Individual sampling point is with second section discrete sample x ₂' ring shift left

Individual sampling point, first section discrete sample x after then obtaining moving to left ₁, second section discrete sample x after moving to left ₂For:

x ₁＝[0.7660 -0.8039 -0.3907 0.9863 0.9703 0.0087 -0.97440.4462]

x ₂＝[0.7660 -0.8039 -0.3907 0.9863 -0.0698 0.0087 -0.97440.4462]

Respectively to x ₁, x ₂Ask DFT, then its DFT result is { X ₁(k), k=0 .., 7}, { X ₂(k), k=0 .., 7}

{X ₁(k)}＝[-0.0315 -0.1207-0.3910i 2.0614+2.2276i 1.7923+0.7763i -1.30621.7923-0.7763i 2.0614-2.2276i -0.1207+0.3910i]

{X ₂(k)}＝[1.0086 -1.1607-0.3910i 3.1014+2.2276i 0.7522+0.7763i -0.26610.7522-0.7763i 3.1014-2.2276i-1.1607+0.3910i]

Obviously, X ₁(k), X ₂(k) the amplitude peak k=2 place that coexists is with X ₁(2), X ₂(2) the value summation can get Y (2)=(2.0614+2.2276i)+(3.1014+2.2276i)=5.1628+4.4552i.

Get the phase angle of Y (2), available phase estimation value

Because said method need carry out twice DFT computing, it is bigger to expend calculated amount, and in order to simplify calculating, the embodiment of the invention has proposed a kind of survey phase control method, and referring to Fig. 5 and Fig. 6, this method may further comprise the steps:

201: set sample frequency f _sWith analysis of spectrum exponent number N;

202: with sample frequency f _sInput simulating signal x (t) is carried out equal interval sampling, obtain

Be total to N+1 discrete sample;

203: with the sample in N+1 the discrete sample

Take advantage of 2, first sampling point x (N/2) with last sampling point x (N/2) addition, is obtained N sampling point y (0)～y (N-1);

$y\left(n\right)=\left\{\begin{array}{c}2x\left(n\right),n\&Element;[0,\frac{N}{2}-1]\\ x(-\frac{N}{2})+x\left(\frac{N}{2}\right),n=\frac{\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="HDA0000036676480000042.png,HDA0000036676480000051.png"\; state="\{\{state\}\}">N</figure-callout>}}{2}\\ 2x(n-N),n\&Element;[\frac{\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="HDA0000036676480000042.png,HDA0000036676480000051.png"\; state="\{\{state\}\}">N</figure-callout>}}{2}+1,N-1]\end{array}\right.$

204: N sampling point y (0)～y (N-1) is DFT, obtains spectrum Y (k);

205:, obtain the amplitude peak of spectrum Y (k) to spectrum Y (k) delivery value;

206: the phase value ang[Y (k*) that will compose the amplitude peak correspondence of Y (k)] as the phase estimation value of center sampling point x (0);

207: judge whether phase estimation value equals second threshold value, if output drives and display module is exported phase estimation value; If not, adjust analysis of spectrum exponent number N, re-execute step 202.

In the practical application, if there is flip-flop in simulating signal, can reduce the precision of measurement, in order further to improve precision, after step 201, before the step 202, this method also comprises: eliminate the flip-flop in the simulating signal for this reason.Wherein, by calculating the mean value of sampling point, just can obtain the flip-flop in the simulating signal.

By above-mentioned steps 201-step 207, only need do DFT one time, just can obtain to compose Y (k), with respect to the method for step 101 to the step 108, calculated amount reduces half, can obtain the result of quarter-phase DFT equally.Will This Before individual element multiply by 2 conducts

Individual sequential value With x (N/2) with the summing value of x (N/2) as intermediate sequence value y (N/2), will

After multiply by 2 conducts

Individual sequential value

, the sequence that obtains at last is

{y(n)}＝[1.5321 -1.6077 -0.7815 1.9726 0.9005 0.0175-1.9487 0.8924]

After { y (n) } be 1 DFT, can get:

{Y(k)}＝[0.9771 -1.2814-0.7819i 5.1628+4.4552i 2.5445+1.5526i-1.5723 2.5445-1.5526i 5.1628-4.4552i -.2814+0.7819i]

The phase angle of getting amplitude peak spectrum Y (the 2)=5.1628+4.4552i of k=2 place can obtain phase estimation value equally

The elaborating of above data processing proved that the quarter-phase DFT among the empty frame of Fig. 2 and Fig. 4 is equivalent, all is only to have used 9 sampling points, promptly realized the accurate estimation to center sampling point x (0) phase place.

Fig. 7 has provided two son spectrum X ₁(k), X ₂(k) with synthetic after the spectrogram (comprising spectral amplitude and phase spectrum) of quarter-phase DFT spectrum Y (k).Can visually see from Fig. 7: | X ₁(k) |, | X ₂(k) | and | Y (k) | the coexist position of k=2 of amplitude peak spectrum, it should be noted that:

Value is 47.2194 degree (being higher than 40 degree), and

Value is 35.6878 degree (being lower than 40 degree), two son spectrum X ₁(k), X ₂(k) phase estimation deviation is positive and negative just the opposite, so two son spectrum X ₁(k), X ₂(k) after the stack, error can positive and negative counteracting and compensation mutually, makes final phase estimation value more close to actual value, reaches

0.7923 ° of this small phase estimation deviate is that two the sidebands spectrum phases mutual interference owing to cosine signal causes.Prove through experimental data: when the N value increased, the interference of two sideband spectrums can reduce, and makes the phase estimation deviate further reduce.

Verify the feasibility of the method that the embodiment of the invention provides below with two concrete examples, see for details hereinafter and describe:

Sampling rate f _s=100Hz is to signal x (t)=cos (2 π f ₀T+ θ ₀), f ₀=20Hz, θ ₀=40 ° of samplings are carried out quarter-phase DFT analysis of spectrum to gathering sample again, and its analysis of spectrum exponent number N from 16 to 68 is with interval 4 incremental variations, the frequency resolution Δ f=f of corresponding D FT _s/ N, table 1 have provided its phase estimation result.

Table 1

Can find out from table 1,, can find following rule along with the N increase (collection is counted and is N+1) of analysis of spectrum exponent number:

1, the phase estimation deviate of the quarter-phase DFT trend that tapers off generally, but successively decreasing of phase estimation deviate is not dull, but relevant with the multiple of signal frequency and frequency resolution Δ f.

2, when N got some particular value (as N=20,40 and 60), its estimated bias value can be decremented to 0, and this is because on these particular values, the frequency resolution f of DFT _s/ N is respectively 5Hz, 2.5Hz and 50/3Hz, and signal frequency f ₀=20Hz is the integral multiple (promptly 4 times, 8 times and 12 times) of these frequency resolutions just, and at this moment DFT does not exist spectrum as shown in Figure 7 to leak, and two sideband spectrums can the phase mutual interference, so its phase estimation deviate is 0.

3, situation about departing from for identical Δ f multiple is got 16,36 and 56 (frequency departure Δ f value is all 0.2) as N, and then the phase estimation deviation is along with exponent number N increases and monotone decreasing.

When 4, the analysis of spectrum exponent number is more or less the same, its phase estimation error is dull the increase with departing from Δ f multiple value to increase, get 20,24 and 28 o'clock (it departs from Δ f multiple value and is respectively 0 ,-0.2 and-0.4) as N, then its error increases to 0.7008 degree from 0 degree, 0.4383 degree.

Sampling rate f _s=100Hz is to signal x (t)=cos (2 π f ₀T+ θ ₀), f ₀=20Hz, θ ₀=40 ° of samplings, carry out 64 rank quarter-phase DFT analysis of spectrums (gathering 65 sampling points) to gathering sample, at this moment noise is subjected to the white noise interference, suppose that its signal to noise ratio (S/N ratio) is SNR (signal to noise ratio, SNR), to every kind of SNR situation, be 1000 Monte Carlo quarter-phase DFT and survey simulation mutually, and counting the standard deviation of this survey phase error of 1000 times, Fig. 8 has provided the survey phase standard deviation curve under the different signal to noise ratio (S/N ratio)s that experiment draws.

As can be seen from Figure 8, surveying the phase standard deviation increases and dull the reduction with signal to noise ratio snr.When SNR＞16dB, quarter-phase DFT surveys the phase standard deviation and can be controlled in 5 degree, meets common engineering application demand.Thereby the method that the embodiment of the invention provides has stronger noise resisting ability.During SNR=50dB, surveying the phase standard deviation only is about 0.13 degree, has very high precision.

In sum, the embodiment of the invention provides a kind of survey to install mutually and control method, and the embodiment of the invention can realize the accurate estimation to center sampling point instantaneous phase with short sample data; Do not need estimated frequency just can realize phase estimation; Owing to do not carry out Frequency Estimation, avoided the error of Frequency Estimation, make that the precision of phase estimation value is very high; After adopting further improvement, only carry out DFT one time, reduced calculated amount, computation complexity is low; Owing to introduced DFT, make that most of noise energy is shared on the very wide non-signal band in the frequency domain, greatly reduce the interference of noise to phase measurement, higher noise resisting ability is arranged; Need not reference signal and can measure phase place, reduced, satisfied the needs in the practical application the cost of hardware and the requirement of complexity.

It will be appreciated by those skilled in the art that accompanying drawing is the synoptic diagram of a preferred embodiment, the invention described above embodiment sequence number is not represented the quality of embodiment just to description.

The above only is preferred embodiment of the present invention, and is in order to restriction the present invention, within the spirit and principles in the present invention not all, any modification of being done, is equal to replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (5)

1. a survey is installed mutually, it is characterized in that described survey device mutually comprises: modulate circuit, A/D converter, digital signal processor and output driving and display circuit thereof,

Simulating signal inserts described modulate circuit, described modulate circuit is adjusted the amplitude of described simulating signal according to first threshold, obtain first simulating signal, described first simulating signal obtains sample sequence through described A/D converter sampling, described sample sequence inserts described digital signal processor with the form of Parallel Digital input, obtain the phase estimation value of sample sequence through the processing of described digital signal processor, the phase estimation value of described sample sequence is by described output driving and display circuit output thereof and show.

2. the control method that a kind of survey according to claim 1 is installed mutually is characterized in that, said method comprising the steps of:

(1) sets sample frequency f _sWith analysis of spectrum exponent number N;

(2) with described sampling frequency f _sInput simulating signal x (t) is carried out equal interval sampling, obtain

Be total to N+1 discrete sample;

(3) a described N+1 discrete sample is divided into two sections, first section discrete sample is by the top n discrete sample

Form, second section discrete sample is by N the discrete sample in back Form;

(4) with described first section discrete sample ring shift left

, obtain first section discrete sample after moving to left; With described second section discrete sample ring shift left Obtain second section discrete sample after moving to left;

(5) first section discrete sample after described the moving to left done the DFT conversion, second section discrete sample after described the moving to left done the DFT conversion,, obtain spectrum Y (k) the summation of the result after the secondary DFT conversion, wherein, Y (k), k=0 ..., N-1};

(6), obtain the amplitude peak of spectrum Y (k) to described spectrum Y (k) delivery value;

(7) with the phase value ang[Y (k*) of the amplitude peak correspondence of described spectrum Y (k)] as the phase estimation value of center sampling point x (0);

(8) judge whether described phase estimation value equals second threshold value, if output drives and display circuit is exported described phase estimation value; If not, adjust described analysis of spectrum exponent number N, re-execute step (2).

3. control method according to claim 2 is characterized in that, in step (1) afterwards, step (2) before, described method also comprises: eliminate the flip-flop among the described simulating signal x (t).

4. the control method that a kind of survey according to claim 1 is installed mutually is characterized in that, said method comprising the steps of:

(9) set sampling f frequently _sWith analysis of spectrum exponent number N;

(10) with described sampling frequency f _sInput simulating signal x (t) is carried out equal interval sampling, obtain

Be total to N+1 discrete sample;

(11) with the sample in the described N+1 discrete sample Take advantage of 2, first sampling point x (N/2) with last sampling point x (N/2) addition, is obtained N sampling point y (0)～y (N-1);

$y\left(n\right)=\left\{\begin{array}{c}2x\left(n\right),n\&Element;[0,\frac{N}{2}-1]\\ x(-\frac{N}{2})+x\left(\frac{N}{2}\right),n=\frac{N}{2}\\ 2x(n-N),n\&Element;[\frac{N}{2}+1,N-1]\end{array}\right.$

(12) described N sampling point y (0)～y (N-1) is done DFT 1 time, obtain spectrum Y (k);

(13), obtain the amplitude peak of spectrum Y (k) to described spectrum Y (k) delivery value;

(14) with the phase value ang[Y (k*) of the amplitude peak correspondence of described spectrum Y (k)] as the phase estimation value of center sampling point x (0);

(15) judge whether described phase estimation value equals second threshold value, if output drives and display circuit is exported described phase estimation value; If not, adjust described analysis of spectrum exponent number N, re-execute step (10).

5. control method according to claim 4 is characterized in that, in step (9) afterwards, step (10) before, described method also comprises: eliminate the flip-flop among the simulating signal x (t).

Images