CN103532906B - Digital modulation error parameter metering method and system based on waveform design - Google Patents

Abstract

The invention discloses a digital modulation error parameter metering method and a system based on waveform design. The metering method comprises the following steps that a, beginning and end continuity of digital modulation signal baseband waveform with finite length is realized; b, setting of a digital modulation error is realized based on error symbol design; and c, an intact error symbol frame is formed. The method and the system achieve metering of digital modulation error parameters such as EVM (Error Vector Magnitude); an experiment indicates that the metering method and the system have the following characteristics that 1, continuous adjustability of a digital modulation error set value within a larger range can be realized; expected digital modulation error parameters can be set; 2, measured values of the digital modulation parameters such as the EVM are highly identical with expected calculated values; 3, readings of the measured values of the digital modulation parameters have very good repeatability; and 4, the digital modulation error parameter metering method and the system are suitable for almost all digital modulation modes, such as 256 QAM (Quadrature Amplitude Modulation) and higher-order modulation modes.

Description

Digital modulation error parameter metering method based on Waveform Design and system

Technical field

The present invention relates to digital communication technology field, integrated wireless communication tester or VSA are carried out Digital modulation error parameter metering method based on Waveform Design and system.

Background technology

Digital modulation signals are the important carriers of present information transmission, it has also become one of foundation stone of information-intensive society, such as no Line communication system just using multiple digital modulation, such as gmsk, qpsk, 16qam, 64qam, 256qam etc..Wherein digital modulation is missed Difference parameter includes Error Vector Magnitude root-mean-square value (hereinafter referred to as evmrms), Error Vector Magnitude peak value (hereinafter referred to as Evmpeak), range error root-mean-square value (hereinafter referred to as magerrrms), range error peak value be (hereinafter referred to as Magerrpeak), phase error root-mean-square value (hereinafter referred to as phaseerrrms), phase error peak value be (hereinafter referred to as Phaseerrpeak) it is the important errors parameter weighing digital modulation signals error.The same with all of physical quantity, these errors Parameter also has to pass through metering that is believable, can tracing to the source, and guarantee value is accurately and unified, and guarantee information system is just Often operate.However, still suffering from following problem in current digital modulation error parameter metering:

1st, closed loop is mutually measured it is difficult to trace to the source.This is primarily referred to as: according to digital modulation error parameter metering side both domestic and external at present Method (such as jjf1128-2004 VSA calibrating standard or jjf1174-2007 digital signal generator calibrating standard The digital modulation error parameter metering method of regulation), on the one hand, digital signal generator is with VSA (below Abbreviation vsa) come to calibrate, that is, each error parameter of the generated digital modulation signals of digital signal generator is not known , but measured and analyzed by vsa and draw；On the other hand, digital signal generator is then based on as standard to the calibration of vsa Device.Above-mentioned metering process is substantially the process that a closed loop is mutually measured, and is the process that a closed loop that can not trace to the source is mutually measured, and is In measuring system " castles in the air ", as shown in Figure 1.That is, the error that current digital modulation error parameter metering obtains Parameter can not be traced to the source, and this is it is difficult to ensure that the accuracy of metering process and credibility.

2nd, the digital modulation signals as calibration vsa standard lack the setting of error parameter.In systems in practice, due to many The interference of the factor of kind, often there is error in digital modulation signals.Therefore, numeral tune can accurately be measured in order to ensure vsa Each error parameter of signal processed, the error parameter as the digital modulation signals of calibration vsa standard within the specific limits should When can arrange.One of effect of vsa is exactly the evm value measuring in relative broad range, and such as td-lte standard requires 16qam The evmrms < 8% of evmrms < the 12.5%.64qam signal of signal, should arrange different evm values at least in the range of this. As the range of certain slide calliper rule is 10 centimetres, and only make the gauge block of 0.1 centimetre of length remove to calibrate slide calliper rule, be inadequate.China is studied Personnel have indicated Similar Problems in pertinent literature.However, the calibration steps of Contemporary Digital modulation error parameter is but not Error parameter is configured, this is an off real requirement.

Content of the invention

For solving technical problem present in above-mentioned prior art, realize the purpose of the present invention, the present invention proposes a kind of base Digital modulation error parameter metering method and system in Waveform Design.

The method comprises the following steps: a, realizes the continuous step of finite length digital modulation signals baseband waveform head and the tail； B, the basis based on error symbol design, the step realizing the setting of digital modulation error；C, form complete error symbolic frame Step.

Wherein, step a refers under the hardware structure of vector modulation signal generator, the iq of one finite duration of design Baseband waveform, is then introduced into the baseband portion of signal generator, and this limited will be held in the way of loop play by signal generator The baseband waveform of continuous time carries out head and the tail and connects, thus it is continuous to realize finite length digital modulation signals baseband waveform head and the tail.Enter One step ground, realizes step a, can be in three kinds of modes below including:

(1) step a comprises the steps:

Step 1: construction symbol sebolic addressing, it is l symbol period that the time domain impulse of baseband filter blocks length, and construction is not filtered The rear l symbol of the front l symbol of ripple symbol sebolic addressing and sequence constitutes symmetrical relationss；

Step 2: pretreatment, according to the baseband waveform sample rate setting and baseband signalling speed, by unfiltered iq symbol Sequence and baseband filter shock response waveform carry out suitable interpolation and process, and make two waveform corresponding time-domain sampling interval phases Deng；

Step 3: convolution, pretreated for step 2 two values sequence is done convolution, participates in the iq symbol of convolution The sampling number of number sequence is n, and corresponding baseband filter shock response waveform sampling points are g, the points of convolution results It is n+g-1；

Step 4: intercept, the convolution results of step 3 are intercepted, intercept the n numerical point in stage casing, intercepting result is Required baseband waveform sequence of values, this sequence has the continuous characteristic of head and the tail.

(2) step a comprises the steps:

Step 1: pretreatment, according to the baseband waveform sample rate setting and baseband signalling speed, by unfiltered iq symbol Sequence and baseband filter shock response waveform carry out suitable interpolation and process and null filling, make the corresponding time domain of two waveforms Sampling interval is equal, and counts and equal be n；

Step 2: fast Fourier transform, pretreated for step 1 two values sequence is in quick Fu of n point respectively Leaf transformation, forms the sequence of two new n ' points: xk1 and xk2, then xk1 with xk2 execution corresponding element order is multiplied one by one Computing, its result yk is also the sequence of values of a n point；

Step 3: an inverse fast fourier, an inverse fast fourier computing is executed to yk, forms the ripple of a time domain Figurate number value sequence s, this sequence is required baseband waveform sequence of values, and this sequence has the continuous characteristic of head and the tail.

(3) step a comprises the steps:

Step 1: sequence repeats, initial iq symbol sebolic addressing comprises m symbol, repeats this sequence, forms new sequence；

Step 2: pretreatment, according to the baseband waveform sample rate setting and baseband signalling speed, by new sequence and base band Filter impulse response wave shape carries out suitable interpolation and processes, or adjustment sample rate, makes the corresponding time-domain sampling of two waveforms Interval is equal；

Step 3: convolution, pretreated for step 1 two values sequence is done convolution computing, participates in convolution The sampling number of iq symbol sebolic addressing is 2n, and corresponding baseband filter shock response waveform sampling points are g, convolution results sequence Represented with ck1, its points are 2n+g-1；

Step 4: intercept, convolution results ck1 of step 2 are intercepted, from selected starting point, continuous n after intercepting The sequence that individual numerical point is constituted, this intercepting result ck2 is required baseband waveform sequence of values, and it is continuously special that this sequence has head and the tail Property, above-mentioned selected starting point in the selection range in ck1 is: from the beginning of the g+1 point, to the n-th -2 points, all optional.

Wherein, step c further includes following concrete steps:

Step 1: construction frame, list the sequence of complex numbers s of one group of standard symbol₁、s₂、s₃…s_k…s_k, for each symbol s_k, According to step b, design corresponding l error symbol s_k1、s_k2…s_kl, form complete error symbolic frame, this symbolic frame be framee；

Step 2: combined frames, according to the requirement of actual test degree symbol sebolic addressing length, by multiple complete error symbolic frames Framee combines, and forms a new sign matrix framem, sign matrix framem restructuring is become the plural number of single file Rearrangement is upset in the position of each symbol then by sign magnitude sequence at random, forms symbol sebolic addressing s_frame；

Step 3: construction waveform, according to step a, based on symbol sebolic addressing s_frameGenerate the continuous baseband signal waveform of head and the tail wb_frameIt is ensured that baseband signal waveform wb_frameIn corresponding symbol sebolic addressing, each symbol that symbolic frame framee comprises is strict Equiprobability occurs；

Step 4: generate rf modulated signal, by baseband waveform wb_frameCarry out proper treatment and be allowed to adaptive vector modulation letter The operation of number generator requires, and baseband waveform is imported the baseband portion of vector modulation signal generator, and then produces radio frequency and adjust Signal wr processed_frame, VSA vsa tested for the input of this radiofrequency signal is demodulated；

Step 5: setting VSA, wb is set_frameCorresponding symbol total number is p, by tested vector letter The symbolic analysis length of number analyser vsa is set to the integral multiple of p.

The present invention also proposes a kind of digital modulation error parameter metering system based on Waveform Design, comprising: for realizing The device of described step a；For realizing the device of described step b；For realizing the device of described step c.

Brief description

A kind of Fig. 1 flow process of the digital modulation error parameter metering method based on Waveform Design provided in an embodiment of the present invention Figure

The metering method of the digital modulation parameter of the closed loop that can not trace to the source in Fig. 2 prior art

Fig. 3 a unfiltered sign magnitude sequence

Fig. 3 b baseband filter shock response waveform

Fig. 4 do not take steps a process baseband filtering after the 16qam signal baseband waveform of finite duration head Tail junction pattern

The corresponding Error Vector Magnitude of each symbol (evm) after the discontinuous waveform loop play of Fig. 5 head and the tail

The construction of the iq symbol sebolic addressing of Fig. 6 the first implementation of step a

The initial part of Fig. 7 iq symbol sebolic addressing

The initial part of the new iq symbol sebolic addressing of Fig. 8 the third implementation of step a

Fig. 9 is using the head and the tail junction of the 64qam signal baseband waveform after the continuous scheme of waveform head and the tail

Figure 10 is using the corresponding error of each symbol after the 64qam baseband signal loop play after the continuous scheme of waveform head and the tail Amplitude of the vector (evm)

Digital modulation error coordinate points setting in the corresponding symbol judgement domain of Figure 11 a point

Axial equal proportion equilateral triangle design in Figure 12 iq plane

The complete error symbol frame of Figure 13 mono-: framee

Figure 14, according to 16qam modulation system, is provided with the vsa demodulated methed vectogram of the digital modulation signals of standard error

Figure 15 is according to expected setting evmrms value (expected evmrms) of 16qam modulation system, value of calculation (calculated) and measurement result (measured) relation;

Wherein, top figure: measure the evmrms obtaining；Middle figure: evmrms measured value and the difference calculating desired value, it is expressed as Measurement error (measurement error)；Base map: the standard deviation (reading time interval 0.1 second) of survey measurementss, in figure evmrms refers to Error Vector Magnitude root-mean-square value；

Figure 16 is according to expected setting magerrrms value (expected magerrrms) of 16qam modulation system, value of calculation (calculated) and measurement result (measured) relation,

Wherein, top figure: measure the magerrrms obtaining；Middle figure: magerrrms measured value and the difference calculating desired value, It is expressed as measurement error (measurement error;Base map: the standard deviation (reading time interval 0.1 second) of survey measurementss, in figure Magerrrms fingerbreadth degree error mean square root;

Figure 17 is according to expected setting phaseerrrms value (expected phaseerrrms) of 16qam modulation system, meter Calculation value (calculated) and the relation of measurement result (measured),

Wherein, top figure: measure the phaseerrrms obtaining；Middle figure: phaseerrrms measured value and calculating desired value Difference, is expressed as measurement error (measurement error)；Base map: the standard deviation (reading time interval 0.1 second) of survey measurementss, In figure phaseerrrms refers to phase error root-mean-square value；

Figure 18 is according to 64qam modulation system, " equilateral triangle radius " γ (linear dimension) and measurement in error setting model The relation of result,

Wherein, top figure: measure the evmrms obtaining;Middle figure: evmrms measured value (measured) and calculating desired value (calculated) difference, is expressed as measurement error (measurement error);Base map: the standard deviation of survey measurementss (is read Number interval 0.1 second), in figure evmrms refers to Error Vector Magnitude root-mean-square value；

Figure 19, according to 64qam modulation system, is provided with the vsa demodulated methed vector of the digital modulation signals of standard error Figure；

Figure 20 is according to 256qam modulation system, " equilateral triangle radius " γ (linear dimension) and measurement in error setting model The relation of result,

Wherein, top figure: measure the evmrms obtaining;Middle figure: evmrms measured value and the difference calculating desired value, it is expressed as Measurement error (measurement error);Base map: the standard deviation (reading time interval 0.1 second) of survey measurementss, in figure evmrms refers to Error Vector Magnitude root-mean-square value;

Figure 21 is according to 256qam modulation system, " equilateral triangle radius " γ (linear dimension) and measurement in error setting model The relation of result,

Wherein, top figure: measure the magerrrms obtaining;Middle figure: magerrrms measured value and the difference calculating desired value, It is expressed as measurement error (measurement error);Base map: the standard deviation (reading time interval 0.1 second) of survey measurementss, in figure Magerrrms fingerbreadth degree error mean square root;

Figure 22 is according to 256qam modulation system, " equilateral triangle radius " γ (linear dimension) and measurement in error setting model The relation of result,

Wherein, top figure: measure the phaseerrrms obtaining;Middle figure: phaseerrrms measured value and calculating desired value Difference, is expressed as measurement error (measurement error);Base map: the standard deviation (reading time interval 0.1 second) of survey measurementss, In figure phaseerrrms refers to phase error root-mean-square value.

Specific embodiment

The embodiment provides a kind of digital modulation error parameter metering method based on Waveform Design.As Fig. 1 Shown, the method comprises the following steps: a, realizes the continuous step of finite length digital modulation signals baseband waveform head and the tail；B, base In the basis of error symbol design, the step realizing the setting of digital modulation error；C, the step forming complete error symbolic frame.

First, finite length digital modulation signals baseband waveform head and the tail are realized continuous

The iq base band of a finite duration under the hardware structure of existing vector modulation signal generator, can be designed Waveform, is then introduced into the baseband portion of signal generator, and signal generator is by when in the way of loop play by this finite-duration Between baseband waveform carry out head and the tail connect, thus constantly modulate carrier wave, formed rf modulated signal.

Form the sequence of values of the iq baseband waveform of a finite duration using the mode of Digital Signal Processing, existing Three steps are typically had in technology:

1st, define the sequence of values of unfiltered symbol (i road and q road define respectively).Shown in (a) of following Fig. 3, in order to It is easy to represent, Fig. 3 (a) illustrate only the sequence on i road, but the symbol sebolic addressing participating in computing comprises i road and q road, typically with multiple The form of number real part and imaginary part represents, no longer individually illustrates below.

2nd, define the time domain impulse response sequence of values of baseband filter, such as conventional raised cosine filter, root liter are remaining String wave filter, Gaussian filter etc..

3rd, the time domain impulse response wave shape of unfiltered sign magnitude sequence and baseband filter does convolution in time domain, realizes filter Ripple, that is, define " the iq baseband waveform of finite duration ".

But in the case of not taking special handling, the waveform that above-mentioned convolution obtains has disadvantages in that waveform sequence Row first and last numerical value are discontinuous, so due to being to connect to come by head and the tail in the internal waveform of signal generator Realize loop play, then the waveform in head and the tail junction will be discontinuous, as shown in Figure 4.

As shown in figure 4, baseband waveform will lead to the extra error of signal in the discontinuous of head and the tail junction, such as measurement figure The Error Vector Magnitude (evm) on every symbol of signal shown in 4, result such as Fig. 5.

Analysis Fig. 5 understands, due to discontinuous in head and the tail junction, leads to head and the tail junction to characterize the evm of signal errors Value significantly increases, and obvious error and uprushes.Obviously, such signal waveform cannot function as calibrating the standard signal of vsa.

Only achieve baseband waveform head and the tail continuously, signal is possible to the metering for digital modulation error parameter, is Realize that baseband waveform head and the tail are continuous, the present invention proposes 3 kinds of technical schemes that can independently use.

Waveform head and the tail continuously realize technical scheme 1:

Step 1: construction symbol sebolic addressing.Frequency domain filtering is equivalent to convolution, and the time domain impulse of such as baseband filter blocks Length is l symbol period, then the rear l symbol of the front l symbol and sequence that construct unfiltered symbol sebolic addressing constitutes symmetrical closing System.As shown in Figure 6.

Step 2: pretreatment.According to the baseband waveform sample rate setting and baseband signalling speed, by unfiltered iq symbol Baseband filter shock response waveform shown in sequence and Fig. 3 (b) carries out suitable interpolation and processes, when making two waveforms corresponding The domain sampling interval is equal.

Step 3: convolution.Two values sequence pretreated described in step 2 is done convolution.If participating in volume The sampling number of long-pending iq symbol sebolic addressing is n, and corresponding baseband filter shock response waveform sampling points are g, then volume The points of long-pending result are n+g-1.

Step 4: intercept.Convolution results described in step 3 are intercepted, only intercepts the n numerical point in stage casing.Should Intercepting result is exactly the final baseband waveform sequence of values needing, and this sequence has the continuous characteristic of head and the tail.

Waveform head and the tail continuously realize technical scheme 2:

Step 1: pretreatment.According to the baseband waveform sample rate setting and baseband signalling speed, by unfiltered iq symbol Baseband filter shock response waveform shown in sequence and Fig. 3 (b) carries out suitable interpolation and processes and null filling, makes two ripples Shape corresponding time-domain sampling interval is equal, and counts and equal be n.

Step 2: fast Fourier transform.Pretreated for step 1 two values sequence is in quick Fu of n point respectively Leaf transformation (fft), this is a kind of digital signal processing algorithm of standard.Define the sequence of two new n points: xk1 and xk2, Then to xk1 and xk2 execution corresponding element order multiplication operation one by one, its result yk is also the sequence of values of a n point.

Step 3: an inverse fast fourier.To yk execution an inverse fast fourier computing (ifft), when forming one It is simply that the final baseband waveform sequence of values needing, this sequence has the continuous characteristic of head and the tail for wave numerics sequence s in domain.

Waveform head and the tail continuously realize technical scheme 3:

Step 1: sequence repeats.Iq symbol sebolic addressing as initial comprises m symbol, can be expressed as form shown in Fig. 6.So After repeat this sequence, form new sequence, as shown in Figure 8.

Step 2: pretreatment.According to the baseband waveform sample rate setting and baseband signalling speed, by the sequence of iq symbol shown in Fig. 8 Row and the baseband filter shock response waveform shown in Fig. 3 (b) carry out suitable interpolation and process, or adjustment sample rate, make two Individual waveform corresponding time-domain sampling interval is equal.

Step 3: convolution.Two values sequence pretreated described in step 2 is done convolution computing.If ginseng Sampling number with the iq symbol sebolic addressing of convolution is 2n, and corresponding baseband filter shock response waveform sampling points are g, that Convolution results sequence can be represented with ck1, and its points are 2n+g-1.

Step 4: intercept.Convolution results ck1 described in step 3 are intercepted, can connect after intercepting from selected starting point The sequence that n continuous numerical point is constituted, this intercepting result ck2 is exactly the final baseband waveform sequence of values needing, and this sequence has There is the continuous characteristic of head and the tail.Above-mentioned selected starting point in the selection range in ck1 is: from the beginning of the g+1 point, to the n-th -2 Point, all optional.

Above three scheme is independent, and is all effective, after the continuous technical scheme of head and the tail waveform, Ke Yida Shown in effect following Fig. 9, the Figure 10 arriving.

Contrast Fig. 9 and Fig. 4, Figure 10 and Fig. 5 is it is clear that have: the continuous scheme of waveform head and the tail is effective, thus determines this One basic technology of invention.

2nd, the digital modulation error setting on the basis based on error symbol design

The symbol sebolic addressing designing digital modulation error it is necessary to (undistorted error free) of establishing criteria goes design error to accord with Number.The sequence of complex numbers assuming the corresponding standard symbol of certain digital modulation mode is s₁、s₂、s₃…s_k…s_kIt is clear that each is multiple Coordinate points in the corresponding iq plane of numerical symbol.Generally, it is assumed that s_kCorresponding a point in iq plane is logical according to classical numeral Letter is theoretical, there is a region and be referred to as " symbol s in iq plane_kJudgement domain ", if the sample vector that obtains of judgement falls into This region, all adjudicates as symbol s_k, as schematic diagram, " symbol s_kJudgement domain " following Figure 11 gray area shown in, then Point beyond several a points of design in judgement domain, these points are labeled as a₁、a₂…a_l, their corresponding complex values are followed successively by s_k1、s_k2…s_kl.So just can count in iq plane, using the definition of the digital modulation error parameter such as geometrical relationship and evm Calculate this setting corresponding digital modulation error parameter, thus realizing the setting of digital modulation error.L >=1, for different symbols Number, l can equal it is also possible to.

Iq coordinate points a₁、a₂…a_lGeometric pattern design can be varied, such as triangle, pentagon, hexagon etc. Deng.The present invention provides one kind and preferably designs example, and this method can be referred to as " axial equal proportion equilateral triangle design ", As shown in figure 12.

In Figure 12, a point is the corresponding iq coordinate points of standard symbol, if a point complex coordinates are: i+jq, vector's Amplitude is m_a, phase angle is θ_a, then adjudicate a of setting in domain in a point₁、a₂、a₃3 points of phasor coordinate is respectively as formula (1)～(3) Shown.

$\stackrel{\&rightarrow;}{{\mathrm{oa}}_1}=m_a(1-\mathrm{\&gamma;})\exp \left({\mathrm{j\&theta;}}_a\right)---\left(1\right)$

$\stackrel{\&rightarrow;}{{\mathrm{oa}}_2}=\stackrel{\&rightarrow;}{{\mathrm{oa}}_1}+\sqrt{3}\mathrm{\&gamma;}{m}_a\exp \left[j({\mathrm{\&theta;}}_a-\frac{\mathrm{\&pi;}}{6})\right]---\left(2\right)$

$\stackrel{\&rightarrow;}{{\mathrm{oa}}_3}=\stackrel{\&rightarrow;}{{\mathrm{oa}}_1}+\sqrt{3}\mathrm{\&gamma;}{m}_a\exp \left[j({\mathrm{\&theta;}}_a+\frac{\mathrm{\&pi;}}{6})\right]---\left(3\right)$

Can the γ in (1)～(3) be called " equilateral triangle radius ", due to the geometrical property of equal proportion, can be by such as Lower simplified operation is calculating corresponding digital modulation error.If the corresponding complex coordinates of a point are 1, then have:

$\stackrel{\&rightarrow;}{{\mathrm{oa}}_1}=1-\mathrm{\&gamma;}---\left(4\right)$

$\stackrel{\&rightarrow;}{{\mathrm{oa}}_2}=1+\frac{\mathrm{\&gamma;}}{2}+j\frac{\sqrt{3}\mathrm{\&gamma;}}{2}---\left(5\right)$

$\stackrel{\&rightarrow;}{{\mathrm{oa}}_3}=1+\frac{\mathrm{\&gamma;}}{2}-j\frac{\sqrt{3}\mathrm{\&gamma;}}{2}---\left(6\right)$

According to symmetry, reference constellation point a ' should be on the extended line of oa, and the computational methods according to root-mean-square value obtain:

$\stackrel{\&rightarrow;}{{\mathrm{oa}}^{\&prime;}}=m_{a^{\&prime;}}=\sqrt{1+{\mathrm{\&gamma;}}^2}---\left(7\right)$

$\mathrm{phaseerrpeak}=\arctan \left(\frac{\frac{\sqrt{3}\mathrm{\&gamma;}}{2}}{1+\frac{\mathrm{\&gamma;}}{2}}\right)---\left(8\right)$

$\mathrm{phaseerrrms}=\sqrt{\frac{{2\mathrm{phaseerrpeak}}^2}{3}}=\sqrt{\frac{2}{3}}\mathrm{phaseerrpeak}---\left(9\right)$

$\mathrm{magerrpeak}=\mathrm{evmpeak}=\frac{\sqrt{{1+\mathrm{\&gamma;}}^2}-(1-\mathrm{\&gamma;})}{\sqrt{1+{\mathrm{\&gamma;}}^2}}---\left(10\right)$

Magerrrms and evmrms carries out vector calculus according to definition.

magerr_i=| m_a'-|oa_i| | i=1,2,3 (11)

$\mathrm{magerrrms}=\frac{\sqrt{\frac{\underset{i=1}{\overset{3}{\mathrm{\&sigma;}}}{\mathrm{magerr}}_i^2}{3}}}{m_{a^{\text{'}}}}\&times;100\%---\left(12\right)$

${\mathrm{evm}}_i=|\stackrel{\&rightarrow;}{{\mathrm{oa}}_i}-\stackrel{\&rightarrow;}{{\mathrm{oa}}^{\&prime;}}|,i=\mathrm{1,2,3}---\left(13\right)$

$\mathrm{evmrms}=\frac{\sqrt{\frac{\underset{i=1}{\overset{3}{\mathrm{\&sigma;}}}{\mathrm{evm}}_i^2}{3}}}{m_{a^{\&prime;}}}\&times;100\%---\left(14\right)$

Formula (8) to (14) gives the functional relationship of modulation error parameter and " equilateral triangle radius " γ.In some works In Cheng Yingyong, give the digital modulation parameters such as the evmrms needing setting it is desirable to corresponding γ-value, that is, need to be similar to (15) formula Functional relationship:

γ=f (evmrms) (15)

For functional relationship shown in (15) formula, contrafunctional method can be analysed it is also possible to count by interpolation etc. by solving Value method is tried to achieve, and can reach very high precision.

3rd, form complete error symbolic frame

This step will ensure that strict equiprobability each error symbol occurs in the measurements, reduces symbol and randomness to metering The adverse effect of test is it is ensured that the strict unification of mathematical model and measurement mechanism.

Specifically comprise the following steps that

1st, construct frame.List the sequence of complex numbers s of certain one group of standard symbol of digital modulation mode₁、s₂、s₃…s_k…s_k, right In each symbol s_k, according to step b, design corresponding l error symbol s_k1、s_k2…s_kl.Then formed complete as shown in fig. 13 that Whole error symbol frame, this symbolic frame is named as framee.As shown in figure 13.

2nd, combined frames.According to the requirement of actual test degree symbol sebolic addressing length, by multiple complete error symbol frames Framee combines, and material is thus formed a new sign matrix framem, framem restructuring is become the plural number of single file Rearrangement is upset in the position of each symbol then by sign magnitude sequence at random, forms symbol sebolic addressing s_frame.

3rd, construct waveform.Using step a, based on s_frameGenerate continuous baseband signal waveform wb of head and the tail_frameIt is ensured that wb_frameIn corresponding symbol sebolic addressing, each symbol that framee comprises is that strict equiprobability occurs.

4th, generate rf modulated signal.By baseband waveform wb_frameCarry out proper treatment and be allowed to adaptive vector modulated signal sending out The operation of raw device requires, and baseband waveform is imported the baseband portion of vector modulation signal generator, and then produces rf modulations letter Number wr_frame, VSA (vsa) tested for the input of this radiofrequency signal is demodulated.

5th, setting VSA (vsa).If wb_frameCorresponding symbol total number is p, then by tested vector The symbolic analysis length of signal analyzer (vsa) is set to the integral multiple of p, this reduces symbol and randomness to metering The adverse effect of test is it is ensured that the mathematical model of digital modulation error setting and measurement mechanism can strictly be unified.

Based on said method and coordinate respective vectors modulation signal generator hardware of the prior art, just construct " base Digital modulation error parameter metering device in head and the tail continuous wave ", if calculate thering is the preferable goodness of fit with actual measurement, and Digital modulation error measuring value has good repeatability and stability, then this device can serve as Transfer Standards or metering Standard uses.

Build experiment test system according to the method described above, produced and there is the digital modulation signals wr being provided with error_frame, Then input a signal into tested VSA (vsa) to measure, shown in following Figure 14～22 of measurement result.

Solve the measurement problem of the digital modulation error parameters such as evm by the method for the present invention and system, experiment shows this Metering method and system have characteristics that 1, can realize digital modulation error arranges value continuously may be used interior in a big way Adjust, expected digital modulation error parameter can be set out.2nd, the measured value of the digital modulation parameter such as evm and expected calculated value It is identical.3rd, the measured value reading of digital modulation parameter has very good repeatability, is embodied in the standard of reading Difference very little.4th, this digital modulation error parameter metering method and system are applied to almost all of digital modulation mode, such as 256qam modulation and its modulation system of higher order.

Described above is only presently preferred embodiments of the present invention it is impossible to limit the interest field of the present invention with this, according to The equal change that scope of the present invention patent is done, still belongs to the scope that the present invention is covered.

Claims (6)

1. a kind of digital modulation error parameter metering method based on Waveform Design is it is characterised in that the method includes following step Rapid:

A, realize the continuous step of finite length digital modulation signals baseband waveform head and the tail；

B, the basis based on error symbol design, the step realizing the setting of digital modulation error；

C, the step forming complete error symbolic frame；

Wherein, step a refers under the hardware structure of vector modulation signal generator, the baseband wave of one finite duration of design Shape, is then introduced into the baseband portion of signal generator, and described vector modulation signal generator was by should in the way of loop play The baseband waveform of finite duration carries out head and the tail and connects, thus realize finite length digital modulation signals baseband waveform head and the tail connecting Continuous.

2. metering method according to claim 1, wherein, step a comprises the steps:

Step 1: construction symbol sebolic addressing, it is l symbol period that the time domain impulse of baseband filter blocks length, constructs unfiltered symbol The rear l symbol of the front l symbol of number sequence and sequence constitutes symmetrical relationss；

Step 2: pretreatment, according to the baseband waveform sample rate setting and baseband signalling speed, by unfiltered symbol sebolic addressing and base Band filter shock response waveform carries out suitable interpolation and processes, and makes the corresponding time-domain sampling interval of two waveforms equal；

Step 3: convolution, pretreated for step 2 two values sequence is done convolution, participates in the iq symbol sequence of convolution The sampling number of row is n, and corresponding baseband filter shock response waveform sampling points are g, and the points of convolution results are n+ g-1；

Step 4: intercept, the convolution results of step 3 are intercepted, intercept the n numerical point in stage casing, it is required for intercepting result Baseband waveform sequence of values, this sequence has the continuous characteristic of head and the tail.

3. metering method according to claim 1, wherein, step a comprises the steps:

Step 1: pretreatment, according to the baseband waveform sample rate setting and baseband signalling speed, by unfiltered iq symbol sebolic addressing Carry out suitable interpolation with baseband filter shock response waveform to process and null filling, make the corresponding time-domain sampling of two waveforms Interval is equal, and counts and equal be n；

Step 2: fast Fourier transform, the fast Fourier that pretreated for step 1 two values sequence is done n point respectively becomes Change, form the sequence that two new n ' put: xk1 and xk2, then xk1 with xk2 is executed with the fortune that corresponding element order is multiplied one by one Calculate, its result yk is also the sequence of values of a n point；

Step 3: an inverse fast fourier, an inverse fast fourier computing is executed to yk, forms the waveform number of a time domain Value sequence s, this sequence is required baseband waveform sequence of values, and this sequence has the continuous characteristic of head and the tail.

4. metering method according to claim 1, wherein, step a comprises the steps:

Step 1: sequence repeats, initial iq symbol sebolic addressing comprises m symbol, repeats this sequence, forms new sequence；

Step 2: pretreatment, according to the baseband waveform sample rate setting and baseband signalling speed, by new sequence and baseband filtering Device shock response waveform carries out suitable interpolation and processes, or adjustment sample rate, makes the corresponding time-domain sampling interval of two waveforms Equal；

Step 3: convolution, pretreated for step 2 two values sequence is done convolution computing, participates in the iq symbol of convolution The sampling number of number sequence is 2n, and corresponding baseband filter shock response waveform sampling points are g, and convolution results sequence is used Ck1 represents, its points are 2n+g-1；

Step 4: intercept, convolution results ck1 of step 3 are intercepted, from selected starting point, continuous n number after intercepting The sequence that value point is constituted, this intercepting result ck2 is required baseband waveform sequence of values, and this sequence has the continuous characteristic of head and the tail, Above-mentioned selected starting point in the selection range in ck1 is: from the beginning of the g+1 point, to the n-th -2 points, all optional.

5. metering method according to claim 1, wherein, step c comprises the steps:

Step 1: construction frame, list the sequence of complex numbers s of one group of standard symbol₁、s₂、s₃…s_k, for each symbol s_k, according to step B, designs corresponding l error symbol s_k1、s_k2…s_kl, form complete error symbolic frame, this complete error symbolic frame named For framee；

Step 2: combined frames, according to the requirement of actual test degree symbol sebolic addressing length, by multiple complete error symbolic frame framee Combine, form a new sign matrix framem, sign matrix framem restructuring is become the complex symbol number of single file Rearrangement is upset in the position of each symbol then by value sequence at random, forms symbol sebolic addressing s_frame；

Step 3: construction waveform, according to step a, based on symbol sebolic addressing s_frameGenerate the continuous baseband waveform wb of head and the tail_frameIt is ensured that Baseband waveform wb_frameIn corresponding symbol sebolic addressing, each symbol that complete error symbolic frame framee comprises is strict equiprobability Occur；

Step 4: generate rf modulated signal, by baseband waveform wb_frameCarry out proper treatment and be allowed to adaptive vector modulated signal sending out The operation of raw device requires, and baseband waveform is imported the baseband portion of vector modulation signal generator, and then produces rf modulations letter Number, VSA vsa tested for the input of this rf modulated signal is demodulated；

Step 5: setting VSA, baseband waveform wb is set_frameCorresponding symbol total number is p, by tested arrow The symbolic analysis length of amount signal analyzer vsa is set to the integral multiple of p.

6. a kind of metering system of the metering method realized described in one of claim 1-5, comprising:

For realizing the device of described step a；

For realizing the device of described step b；

For realizing the device of described step c.