CN101789835B - Method and device for measuring spectrum occupation ratio - Google Patents

Abstract

The invention discloses a method for measuring a spectrum occupation ratio, which comprises the following steps: detecting signal strength in a channel, and comparing the detected signal strength with a threshold to obtain a comparison result sequence; and according to the law of large numbers and central limit theorem, analyzing the comparison result sequence by following the normal distribution law, and estimating the spectrum occupation ratio, a confidence interval and relative accuracy thereof. The invention also discloses a device for measuring the spectrum occupation ratio. The method and the device can ensure that the measured spectrum occupation ratio has a reasonable confidence interval; besides, under the condition of the same confidence level and relative accuracy, the number of sampling points is much less than that of sampling points in the prior art, particularly under the condition of a high occupation ratio.

Description

The method of measurement of frequency spectrum occupancy rate and device

Technical field

The present invention relates to communication technical field, relate in particular to method of measurement and the device of frequency spectrum occupancy rate.

Background technology

Radio Spectrum Resource is non-renewable valuable source, is one of the key factor that should consider in effective utilization of spectrum management intermediate frequency spectrum.Frequency spectrum occupancy rate is the embodiment of rational spectrum utilization, the measurement result of frequency spectrum occupancy rate not only can provide the information about the actual service condition of frequency spectrum for spectrum management personnel, facilitate spectrum management personnel assigned frequency, the information of frequency spectrum usage trend can also be provided for frequency authorities simultaneously.Frequency spectrum occupancy rate measurement is an important content in daily monitoring radio-frequency spectrum work.

At present, have much about the method for measurement of frequency spectrum occupancy rate, most method of measurement is in long Fundamentals of Measurement, to obtain the numerical value of degree of taking.But these method of measurement have two problems not consider, one, the credibility of degree of taking, i.e. confidential interval, the reliability of measured degree of taking cannot judge; Two, the levels of precision of degree of taking, i.e. relative accuracy.Concerning actual monitoring work, the meaning of degree of taking is guaranteeing certain confidence level in the situation that, how goes to obtain frequency occupation information as much as possible with the least possible work.According to Principle of Statistics, International Telecommunication Association (ITU) has proposed the guidance method of measuring about degree of taking, the concept of clear and definite credibility and levels of precision.Confidential interval, relative accuracy can allow tester know the reliable results degree of test, guaranteeing, under the condition of certain confidence level and relative accuracy, can greatly to shorten Measuring Time, improve the utilization benefit of equipment.

Inventor realizing in process of the present invention, finds that prior art exists following not enough:

In actual measurement, the method of measurement of ITU adopts the Mathematical Modeling of Poisson distribution to carry out frequency spectrum occupancy rate measurement, there is the method for measurement of certain limitation: ITU to be applicable to the situation that degree of taking is little, for the large situation of degree of taking, the confidential interval upper limit there will be the unreasonable situation that is greater than 100%, and the quantity of sampled point is more, make measuring process comparatively complicated cumbersome.

Summary of the invention

The embodiment of the present invention provides a kind of method of measurement of frequency spectrum occupancy rate, in order to guarantee that measured frequency spectrum occupancy rate has rational confidential interval, under the condition of identical confidence level and relative accuracy, reduces the quantity of sampled point, and the method comprises:

Detect signal strength in channel, and the signal strength detecting and thresholding are carried out to size relatively, obtain comparison result sequence;

According to large several central-limit theorems, described comparison result sequence is analyzed to estimated spectral degree of taking and confidential interval thereof and relative accuracy according to Normal Distribution rule.

Preferably, described comparison result sequence is random process X (n):

X (n)=X _i, X (n) obeys the 0-1 that parameter is ρ and distributes, and the probability density function of X (n) is:

F (x, ρ)=ρ ^x(1-ρ) ^1-x, wherein x=0,1,0 < ρ < 1, n is positive integer.

Preferably, according to large several central-limit theorems, described comparison result sequence is analyzed according to Normal Distribution rule, estimated spectral degree of taking, comprising:

According to large several central-limit theorems, determine: $\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{X}_i-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}=\frac{n\stackrel{\&OverBar;}{X}-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}$ The approximate N (0,1) that obeys; Wherein,

represent sample average, the i.e. estimated value of frequency spectrum occupancy rate.

Preferably, according to large several central-limit theorems, described comparison result sequence is analyzed according to Normal Distribution rule, the confidential interval of estimated spectral degree of taking, comprising:

Determine: $P\{-{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\&le;\frac{n\stackrel{\&OverBar;}{X}-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}\&le;{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\}=1-\mathrm{\&alpha;},$ Wherein it is the normal distribution that average is 0, variance is 1

the upper boundary values at place; 1-α is confidence level;

Be equivalent to $(n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2){\mathrm{\&rho;}}^2-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2)\mathrm{\&rho;}+n{\stackrel{\&OverBar;}{X}}^2\&le;0$

If $a=n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2,$ $b=-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2),$ $c=n{\stackrel{\&OverBar;}{X}}^2,$

B ²-4ac>=0, obtains A≤ρ≤B, and wherein [A, B] is the interval estimation of ρ;

According to radical formula, obtain: $A=\frac{-b-\sqrt{b^2-4\mathrm{ac}}}{2a};$ $B=\frac{-b+\sqrt{b^2-4\mathrm{ac}}}{2a}.$

Preferably, according to large several central-limit theorems, described comparison result sequence is analyzed according to Normal Distribution rule, the relative accuracy of estimated spectral degree of taking, comprising:

Obtain relative accuracy Y: $Y=\frac{B-A}{2\stackrel{\&OverBar;}{X}}.$

The embodiment of the present invention also provides a kind of measurement mechanism of frequency spectrum occupancy rate, in order to guarantee that measured frequency spectrum occupancy rate has rational confidential interval, under the condition of identical confidence level and relative accuracy, reduce the quantity of sampled point, particularly, for situation about taking greatly, this device comprises:

Detection module, for detection of signal strength in channel, and carries out size relatively by the signal strength detecting and thresholding, obtains comparison result sequence;

Estimation module, for according to large number central-limit theorem, analyzes described comparison result sequence estimated spectral degree of taking and confidential interval thereof and relative accuracy according to Normal Distribution rule.

Preferably, the comparison result sequence that described detection module obtains is random process X (n):

X (n)=X _i, X (n) obeys the 0-1 that parameter is ρ and distributes, and the probability density function of X (n) is:

F (x, ρ)=ρ ^x(1-ρ) ^1-x, wherein x=0,1,0 < ρ < 1, n is positive integer.

Preferably, described estimation module specifically for:

According to large several central-limit theorems, determine: $\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{X}_i-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}=\frac{n\stackrel{\&OverBar;}{X}-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}$ The approximate N (0,1) that obeys; Wherein,

represent sample average, the i.e. estimated value of frequency spectrum occupancy rate.

Preferably, described estimation module specifically for:

Determine: $P\{-{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\&le;\frac{n\stackrel{\&OverBar;}{X}-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}\&le;{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\}=1-\mathrm{\&alpha;},$ Wherein

it is the normal distribution that average is 0, variance is 1

the upper boundary values at place; 1-α is confidence level;

Be equivalent to $(n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2){\mathrm{\&rho;}}^2-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2)\mathrm{\&rho;}+n{\stackrel{\&OverBar;}{X}}^2\&le;0$

If $a=n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2,$ $b=-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2),$ $c=n{\stackrel{\&OverBar;}{X}}^2,$

B ²-4ac>=0, obtains A≤ρ≤B, and wherein [A, B] is the interval estimation of ρ;

According to radical formula, obtain: $A=\frac{-b-\sqrt{b^2-4\mathrm{ac}}}{2a};$ $B=\frac{-b+\sqrt{b^2-4\mathrm{ac}}}{2a}.$

Preferably, described estimation module specifically for:

Obtain relative accuracy Y: $Y=\frac{B-A}{2\stackrel{\&OverBar;}{X}}.$

In the embodiment of the present invention, detect signal strength in channel, and the signal strength detecting and thresholding are carried out to size relatively, obtain comparison result sequence; According to large several central-limit theorems, described comparison result sequence is analyzed to estimated spectral degree of taking and confidential interval thereof and relative accuracy according to Normal Distribution rule; The Mathematical Modeling that is different from available technology adopting Poisson distribution is carried out frequency spectrum occupancy rate measurement, can guarantee that measured frequency spectrum occupancy rate has rational confidential interval, and, under the condition of identical confidence level and relative accuracy, sampled point quantity is few more a lot of than the sampled point quantity of prior art, particularly for situation about taking greatly.

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, to the accompanying drawing of required use in embodiment or description of the Prior Art be briefly described below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skills, do not paying under the prerequisite of creative work, can also obtain according to these accompanying drawings other accompanying drawing.In the accompanying drawings:

Fig. 1 is the flow chart of the method for measurement of embodiment of the present invention intermediate frequency spectrum degree of taking;

Fig. 2 is the schematic diagram of random process in the embodiment of the present invention;

Fig. 3 is the situation schematic diagram that in the embodiment of the present invention, relative accuracy changes with sampled point;

Fig. 4 be in the embodiment of the present invention degree of taking and sampled point be related to schematic diagram;

Fig. 5 is the structural representation of the measurement mechanism of embodiment of the present invention intermediate frequency spectrum degree of taking.

Embodiment

For making object, technical scheme and the advantage of the embodiment of the present invention clearer, below in conjunction with accompanying drawing, the embodiment of the present invention is described in further details.At this, schematic description and description of the present invention is used for explaining the present invention, but not as a limitation of the invention.

For clearly demonstrating embodiment of the present invention method, the statistics intension of first brief description frequency spectrum occupancy rate:

Degree of taking is a statistical value or perhaps estimated value.In monitoring time, signal strength exceedes a certain given thresholding, is called and takies; Otherwise it is vacant that signal strength is called lower than thresholding, so the transmitting of given channel takies the random process that situation is 2 states.Because channel status is random, thus cannot determine the state in any given moment, but channel status can be described with certain probability.

Some channel a period of times of continuous measurement, the percentage of time taking degree of taking exactly.If a certain channel of continuous measurement one hour, calculating degree of taking is 25%, is accurately so for this channel degree of taking of 25% in this hour; If only done the observation (or measure) of limited number of time within this hour, just can not entirely accurate provide degree of taking, but can remove estimation degree of taking.From statistical angle, be to totally estimating by limited sample to the description of degree of taking.

According to the statistics intension of above-mentioned frequency spectrum occupancy rate, the embodiment of the present invention provides a kind of method of measurement of frequency spectrum occupancy rate, and the method is different from the Mathematical Modeling of available technology adopting Poisson distribution and carries out frequency spectrum occupancy rate measurement, but as shown in Figure 1, comprising:

Signal strength in step 101, detection channel, and the signal strength detecting and thresholding are carried out to size relatively, obtain comparison result sequence;

Step 103, the large number of basis central-limit theorem, analyze described comparison result sequence estimated spectral degree of taking and confidential interval thereof and relative accuracy according to Normal Distribution rule.

Flow process can be learnt as shown in Figure 1, in the embodiment of the present invention, obtaining after comparison result sequence, according to large number central-limit theorem, described comparison result sequence is analyzed according to Normal Distribution rule, estimated spectral degree of taking and confidential interval thereof and relative accuracy, compared with prior art, can make measured frequency spectrum occupancy rate there is rational confidential interval, and, under the condition of identical confidence level and relative accuracy, sampled point quantity is few more a lot of than the sampled point quantity of prior art, particularly for situation about taking greatly.To be described in further detail this below.

As shown in Figure 2, definition stochastic variable V and W.V is launch time, is in monitoring time T, and signal strength exceedes the duration of thresholding Li continuously.W is the time interval between transmitting.T is monitoring time.Suppose the mean value of T much larger than V and W.

Definition random process X (t) is the random process of 2 values, and this random process X (t) is the comparative result of signal strength described in Fig. 1 and thresholding:

X (t)=1, represents that signal strength exceedes thresholding Li in the t moment;

X (t)=0, represents that signal strength does not exceed thresholding Li in the t moment.

Therefore, stochastic variable V can be regarded as random process X (t) and gets continuously time of 1; W can be regarded as random process X (t) and gets continuously time of 0.

Here define mark: β (L _i, T), be illustrated in the T time, X (t), in the time of getting 1, accounts for the ratio of total time T.Here β (L _i, T) and be exactly degree of taking.Because W and V are stochastic variables, so β (L _i, T) and be also a stochastic variable, and be random variable of continuous type.

Suppose that stochastic variable W and V are separate, V+W has continuous distribution.

So according to limit theorem: $\underset{T=\&infin;}{\lim }=P[X\left(t\right)=1]=\frac{E\left(V\right)}{E\left(V\right)+E\left(W\right)}---\left(1\right)$

And within 0 to T moment, β (L _i, T) and meet progressive normal distribution, so

$\mathrm{lim\; P}[\sqrt{T}\frac{\mathrm{\&beta;}(L_i,T)-\mathrm{\&mu;}}{\mathrm{\&sigma;}}\&le;X]=\frac{1}{\sqrt{2\mathrm{\&pi;}}}{\&Integral;}_{-\&infin;}^X{e}^{-\frac{y^2}{2}}\mathrm{dy}---\left(2\right)$

Here, μ is stochastic variable β (L _i, T) mathematic expectaion, the namely mean value of degree of taking; σ is β (L _i, T) variance.Can prove:

$\mathrm{\&mu;}=\frac{E\left(V\right)}{E\left(V\right)+E\left(W\right)}---\left(3\right)$

${\mathrm{\&sigma;}}^2=D\left(X\right)=\frac{E^2\left(V\right)D\left(W\right)+E^2\left(W\right)D\left(V\right)}{{\left\{E\right[V]+E[W\left]\right\}}^3}---\left(4\right)$

Here, E (V) represents the mathematic expectaion of stochastic variable V, and D (V) represents the variance of stochastic variable V.

As can be seen here, a GPRS μ and σ, stochastic variable β (L so _i, T) distribution also just determined.And μ=P[X (t)=1]=ρ, ρ gets 1 probability, and namely signal exceedes the probability of thresholding.

Therefore,, by defining 2 value random process X (t), degree of taking problem is converted into the estimation problem to ρ value.Measurement result is a series of 0-1 sequence, separate between measurement result, and this belongs to Bernoulli Jacob (Bernoulli) test.

In reality test, the measurement of frequency spectrum occupancy rate is completed by receiver rapid scanning.Be equivalent to X (t) to sample, then calculate β (L _i, T), detect signal strength in channel in implementation step 101, and the signal strength detecting and thresholding are carried out to size relatively, obtain comparison result sequence; This comparison result sequence can be designated as random process X (n):

Make X (n)=X _i, comparison result sequence X ₁, X ₂, X ₃... X _ican be expressed as the sequence of 0 and 1 composition.Here in 1 channel that represents to detect, signal strength exceedes thresholding, and in 0 channel that represents to detect, signal strength does not exceed thresholding.

For comparative illustration embodiment of the present invention method and beneficial effect thereof, the method for measurement of ITU in prior art and embodiment of the present invention method are carried out to comparative illustration below:

One, estimated spectral degree of taking:

In the ITU method of measurement of prior art:

Suppose that the occupied number of times of channel is c in n test.The maximal possibility estimation of degree of taking is

$\frac{c}{n}---\left(5\right)$

In the method for measurement of the embodiment of the present invention:

Comparison result sequence X (n) obeys the 0-1 that parameter is ρ and distributes, and the probability density function of X (n) is:

F (x, ρ)=ρ ^x(1-ρ) ^1-x, wherein x=0,1,0 < ρ < 1, n is positive integer.

According to large several central-limit theorems, determine: $\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{X}_i-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}=\frac{n\stackrel{\&OverBar;}{X}-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}$ The approximate N (0,1) that obeys; Wherein,

represent sample average, the i.e. estimated value of frequency spectrum occupancy rate.

Two, the confidential interval of estimated spectral degree of taking, to determine the credibility of frequency spectrum occupancy rate:

In the ITU method of measurement of prior art:

For scanning each time, the occupied number of times of channel is obeyed Bernoulli Jacob and is distributed; In M scanning, the occupied number of times m of channel obeys binomial distribution:

$P(X=m)=C_M^m{\mathrm{\&rho;}}^m{(1-\mathrm{\&rho;})}^{M-m}---\left(6\right)$

So in M scanning, at least detect that the occupied probability once of channel is:

P＝1-P(X＝0)＝1-(1-ρ) ^M (7)

So $M=\frac{\log [1-P]}{\log [1-\mathrm{\&rho;}]}---\left(8\right)$

Here P is the confidence level of requirement.

Make P=0.99, just can obtain under 99% confidence level, the occupied needed measurement of judgement allocated channel is counted, in table one.Certainly, these measurements are independently.

Table one

Channel occupancy degree (ρ × 100) %	The M that counts needing under 99% confidence level condition
		50	7
20	21
		10	44
1	459
		0.1	4603

0.01

46050

Table one illustrates, if done 4603 measurements at given channel, but signal do not detected.Can believe that this channel occupancy degree is below 0.1% with 99% confidence level so.

Due under the certain condition of monitoring time (being that number of test points is certain), ρ value is less, and the estimation of ρ is got over to inaccuracy, so adopt the method for confidential interval to analyze the very little situation of degree of taking.That is to say the approximation that not only will know ρ, also will know the levels of precision that ρ is estimated.Here it should be noted that:

What in prior art, discuss is small probability event, and ρ≤0.1 has here had this prerequisite could use the conclusion of Poisson distribution;

Confidential interval is one group of random interval, no matter what value ρ gets, this interval all comprises ρ with certain probability (being called again confidence level).Here confidence level is 1-α.

Make U and L be respectively the upper and lower fiducial limit of Poisson distribution, utilize known formula to have:

$U=c+\frac{1}{2}+\frac{3}{8}{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}^2+{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\sqrt{c+\frac{1}{2}+\frac{1}{8}{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2}$

$L=c-\frac{1}{2}+\frac{3}{8}{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2-{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\sqrt{c-\frac{1}{2}+\frac{1}{8}{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}^2}---\left(9\right)$

for the normal distribution that average is 0, variance is 1

the upper boundary values at place.

And the upper and lower borderline value of normal distribution that average is 0, variance is 1 can be tabled look-up and two obtained:

Table two

Make again p _uand p _lbe respectively the upper and lower fiducial limit of p, utilize known formula to have:

$p_U=\frac{U}{n+(U-c)/2}$

$p_L=\frac{L}{n-(c-1-L)/2}---\left(10\right)$

Give an example, suppose to have measured 4000 times, wherein have and detect that channel is occupied for 80 times, be i.e. n=4000, c=80.

The maximal possibility estimation of channel occupancy degree is $\widehat{\mathrm{\&rho;}}=\frac{80}{4000}=0.02$

Make 1-α=0.9, setting confidence level is 90%, $\frac{\mathrm{\&alpha;}}{2}=0.05.$ The fiducial limit table of looking into N (0,1) normal distribution, can obtain U=96.3, and L=65.8, so can obtain according to formula (14): p _u=0.0240, p _l=0.0165.

So can reach a conclusion: the confidence level with 90% judges that channel occupancy degree is between 1.65% to 2.4%.

And in the method for measurement of the embodiment of the present invention:

In the large number of basis central-limit theorem, determine: $\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{X}_i-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}=\frac{n\stackrel{\&OverBar;}{X}-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}$ Approximate obedience after N (0,1), confidential interval that can estimated spectral degree of taking, comprises definite: $P\{-{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\&le;\frac{n\stackrel{\&OverBar;}{X}-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}\&le;{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\}=1-\mathrm{\&alpha;},$ Wherein

it is the normal distribution that average is 0, variance is 1 the upper boundary values at place; 1-α is confidence level;

Be equivalent to $(n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2){\mathrm{\&rho;}}^2-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2)\mathrm{\&rho;}+n{\stackrel{\&OverBar;}{X}}^2\&le;0$

If $a=n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2,$ $b=-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2),$ $c=n{\stackrel{\&OverBar;}{X}}^2,$

B ²-4ac>=0, obtains A≤ρ≤B, and wherein [A, B] is the interval estimation of ρ;

According to radical formula, obtain: $A=\frac{-b-\sqrt{b^2-4\mathrm{ac}}}{2a};$ $B=\frac{-b+\sqrt{b^2-4\mathrm{ac}}}{2a}.$

Three, above-described embodiment has provided confidential interval, and this is an absolute magnitude, and in actual measurement, relative precision also needs to consider.The relative accuracy of estimated spectral degree of taking below, to determine the levels of precision of frequency spectrum occupancy rate:

In prior art in the method for measurement of ITU:

Relative accuracy can be defined as: $\frac{U-L}{2c}\&times;100\%---\left(11\right)$

Relative accuracy can be regarded as the relative levels of precision under given confidence level condition.For example above, confidence level is 90%, and c=80, so $\frac{U-L}{2c}\&times;100\%=\frac{30.5}{160}\&times;100\%=19\%,$ So relative accuracy is 19%.If c=8, (U-L)/16=0.656, relative accuracy is now 66%.

In the suggestion and " spectrum monitoring handbook " of ITU, be all to have stipulated under confidence level and relative accuracy condition, provide and measure the suggestion of counting.For example: under 95% confidence level condition, if reach 10% relative accuracy, can calculate c=390 according to formula, that is to say that measuring the occupied number of times of channel will reach 390 times, can calculate n in conjunction with channel occupancy degree, i.e. total the counting of required test.If channel occupancy degree is 10%, n=c/ ρ=3900, need 3900 independent point; If channel occupancy degree is 15%, n=c/ ρ=2600, need 2600 independent sample points.

And in the method for measurement of the embodiment of the present invention:

If receiver scanning survey n time, detect m time occupied, known under 95% confidence level condition, ${\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}=1.96,$ And $\stackrel{\&OverBar;}{X}=\frac{\underset{i=0}{\overset{n}{\mathrm{\&Sigma;}}}{X}_i}{n}=\frac{m}{n}$

Relative accuracy Y: $Y=\frac{B-A}{2\stackrel{\&OverBar;}{X}}$

If $\stackrel{\&OverBar;}{X}=\frac{m}{n}$ Known, degree of taking is known, and the situation that relative accuracy changes with sampled point as shown in Figure 3.If given relative accuracy is 10%, under 95% confidence level condition,

$({4Y}^2{n}^2+{4Y}^2{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^4+{8Y}^{2}n{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2+4n{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2)m^2-{4n}^2{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^{2}m-n^2{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^4=0$

Fig. 4 shows under 10% relative accuracy and 95% confidence level, the relation between degree of taking and sampled point.As seen from Figure 4, under 10% relative accuracy and 95% confidence level condition, degree of taking and the independent sample that the needs relation between counting, as table three:

Table three

Degree of taking (%)	The independent sample point needing	The independent sample point that ITU-R suggestion .185-2 requires
			6.67	5368	5850
10	3461	3900
			15	2177	2600
20	1535	1950
			30	894	1300
40	573	975
			50	381	780
60	253	650
			70	162	557

Contrast table three is distinguishing with ITU-R suggestion .185-2, the reason of this difference is: what ITU-R suggestion .185-2 adopted is the Mathematical Modeling of Poisson distribution, and what in the embodiment of the present invention, adopt according to large number central-limit theorem is the Mathematical Modeling of normal distribution.Because Poisson distribution is used for describing small probability event, i.e. the very little situation of ρ, also i.e. degree of taking

very little situation.Also can find out from table three: under the lower condition of degree of taking, the quantity of independent sample point is more or less the same, and along with the increase of degree of taking, the quantity variance of independent sample point is increasing, adopts point so same for dependent.Under the condition of identical confidence level and relative accuracy, in the embodiment of the present invention, sampled point quantity is few more a lot of than the sampled point quantity of prior art, particularly for situation about taking greatly.

In real work, before degree of taking test, some parameters can also be set, for example, first can estimate to launch duration, then the flyback time that can accept according to the setting of transmitting duration, also can consider to arrange the bin width and the scanning step (being the width of channel) that need monitoring simultaneously.Next threshold level can be set, guarantee that receiving system has under the condition of natural sensitivity, threshold level can arrange lowly as much as possible, but will note avoiding recording noise.Particularly, for the many situations of small-signal, thresholding has arranged high and has easily missed signal, and thresholding has arranged low and easily noise added, and introduces measure error.The information that can also record in addition, has: test position, flyback time, testing time, frequency and take number of degrees value etc.

Test can have high selectivity with the intermediate-frequency filter of receiver, and its form factor can be 2: 1 or better.For the requirement in speed, also can use Gaussian filter.In order not miss signal, arranging of resolution bandwidth can be more bigger than scanning step conventionally.

One of ordinary skill in the art will appreciate that all or part of step realizing in above-described embodiment method is can carry out the hardware that instruction is relevant by program to complete, described program can be stored in a computer read/write memory medium, this program is in the time carrying out, can comprise all or part of step in above-described embodiment method, described storage medium can comprise: ROM, RAM, disk, CD etc.

In the embodiment of the present invention, also provide a kind of measurement mechanism of frequency spectrum occupancy rate, as described in the following examples.Because the principle that this device is dealt with problems is similar to the method for measurement of frequency spectrum occupancy rate, therefore the enforcement of this device can be referring to the enforcement of method, repeats part and repeat no more.

As shown in Figure 5, the measurement mechanism of embodiment of the present invention intermediate frequency spectrum degree of taking can comprise:

Detection module 501, for detection of signal strength in channel, and carries out size relatively by the signal strength detecting and thresholding, obtains comparison result sequence;

Estimation module 502, for according to large number central-limit theorem, analyzes described comparison result sequence estimated spectral degree of taking and confidential interval thereof and relative accuracy according to Normal Distribution rule.

In an embodiment, the comparison result sequence that detection module 501 obtains is random process X (n):

X (n)=X _i, X (n) obeys the 0-1 that parameter is ρ and distributes, and the probability density function of X (n) is:

F (x, ρ)=ρ ^x(1-ρ) ^1-x, wherein x=0,1,0 < ρ < 1, n is positive integer.

In an embodiment, estimation module 502 specifically can be used for:

According to large several central-limit theorems, determine: $\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{X}_i-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}=\frac{n\stackrel{\&OverBar;}{X}-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}$ The approximate N (0,1) that obeys; Wherein,

represent sample average, the i.e. estimated value of frequency spectrum occupancy rate.

In an embodiment, estimation module 502 specifically can be for:

Determine: $P\{-{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\&le;\frac{n\stackrel{\&OverBar;}{X}-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}\&le;{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\}=1-\mathrm{\&alpha;},$ Wherein

it is the normal distribution that average is 0, variance is 1

the upper boundary values at place; 1-α is confidence level;

Be equivalent to $(n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2){\mathrm{\&rho;}}^2-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2)\mathrm{\&rho;}+n{\stackrel{\&OverBar;}{X}}^2\&le;0$

If $a=n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2,$ $b=-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2),$ $c=n{\stackrel{\&OverBar;}{X}}^2,$

B ²-4ac>=0, obtains A≤ρ≤B, and wherein [A, B] is the interval estimation of ρ;

According to radical formula, obtain: $A=\frac{-b-\sqrt{b^2-4\mathrm{ac}}}{2a};$ $B=\frac{-b+\sqrt{b^2-4\mathrm{ac}}}{2a}.$

In an embodiment, estimation module 503 specifically can be for:

Obtain relative accuracy Y: $Y=\frac{B-A}{2\stackrel{\&OverBar;}{X}}.$

In sum, in the embodiment of the present invention, detect signal strength in channel, and the signal strength detecting and thresholding are carried out to size relatively, obtain comparison result sequence; According to large several central-limit theorems, described comparison result sequence is analyzed according to Normal Distribution rule, estimate frequency spectrum occupancy rate; The Mathematical Modeling that is different from available technology adopting Poisson distribution is carried out frequency spectrum occupancy rate measurement, can guarantee that measured frequency spectrum occupancy rate has rational confidential interval.And in the embodiment of the present invention, under the condition of identical confidence level and relative accuracy, sampled point quantity is few more a lot of than the sampled point quantity of prior art, particularly for situation about taking greatly, has simplified the measuring process of frequency spectrum occupancy rate.

Above-described specific embodiment; object of the present invention, technical scheme and beneficial effect are further described; institute is understood that; the foregoing is only specific embodiments of the invention; the protection range being not intended to limit the present invention; within the spirit and principles in the present invention all, any modification of making, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.

Claims (2)

1. a method of measurement for frequency spectrum occupancy rate, is characterized in that, the method guarantees to have rational confidential interval for measured frequency spectrum occupancy rate in degree of taking situation greatly, and the method comprises:

Detect signal strength in channel, and the signal strength detecting and thresholding carried out to size relatively, obtain comparison result sequence, wherein further comprise:

Definition random process X (t) is the random process of 2 values, and X (t)=1 represents that signal strength exceedes thresholding Li in the t moment, and X (t)=0 represents that signal strength does not exceed thresholding Li in the t moment;

X (t) is sampled, and the signal strength detecting and thresholding are carried out to size relatively, obtain comparison result sequence, this comparison result sequence is designated as random process X(n), X(n)=X _i, X(n) to obey parameter be ρ, and 0-1 distributes, X(n) probability density function be:

F (x, ρ)=ρ ^x(1-ρ) ^1-x, wherein x=0,1,0< ρ <1, n is positive integer;

According to large several central-limit theorems, described comparison result sequence is analyzed to estimated spectral degree of taking and confidential interval thereof and relative accuracy according to Normal Distribution rule;

Wherein further comprise:

Definition stochastic variable V and W, V is launch time, and, in monitoring time T, signal strength exceedes the duration of thresholding Li continuously, and W is the time interval between transmitting, and T is much larger than the mean value of V and W;

Definition frequency spectrum occupancy rate β (L _i, T), be illustrated in monitoring time T, X (t) accounts for the ratio of total time, wherein β (L in the time of getting 1 _i, T) and be random variable of continuous type, suppose that stochastic variable W and V are separate, V+W has continuous distribution;

So according to limit theorem: $\underset{T=\&infin;}{\lim }=P[X\left(t\right)=1]=\frac{E\left(V\right)}{E\left(V\right)+E\left(W\right)},$ And within 0 to T moment, β (L _i, T) and meet progressive normal distribution, so $\mathrm{limP}[\sqrt{T}\frac{\mathrm{\&beta;}(L_i,T)-\mathrm{\&mu;}}{\mathrm{\&sigma;}}\&le;X]=\frac{1}{\sqrt{2\mathrm{\&pi;}}}{\&Integral;}_{-\&infin;}^X{e}^{-\frac{y^2}{2}}\mathrm{dy},$ Here, μ is stochastic variable β (L _i, T) mathematic expectaion;

Wherein, estimated spectral degree of taking, comprising:

According to large several central-limit theorems, determine:

approximate N(0,1 of obeying); Wherein,

represent sample average, the i.e. estimated value of frequency spectrum occupancy rate;

The confidential interval of estimated spectral degree of taking, comprising:

According to large several central-limit theorems, determine: $P\{-{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\&le;\frac{n\stackrel{\&OverBar;}{X}-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}\&le;{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\}=1-\mathrm{\&alpha;},$ Wherein

it is the normal distribution that average is 0, variance is 1

the upper boundary values at place; 1-α is confidence level;

Be equivalent to $(n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2){\mathrm{\&rho;}}^2-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2)\mathrm{\&rho;}+n{\stackrel{\&OverBar;}{X}}^2\&le;0$

If $a=n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2,b=-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2),c=n{\stackrel{\&OverBar;}{X}}^2,$

B ²-4ac>=0, obtains A≤ρ≤B, and wherein [A, B] is the interval estimation of ρ;

According to radical formula, obtain: $A=\frac{-b-\sqrt{b^2-4\mathrm{ac}}}{2a};B=\frac{-b+\sqrt{b^2-4\mathrm{ac}}}{2a};$

The relative accuracy that estimates frequency spectrum occupancy rate, comprising:

Obtain relative accuracy Y:

2. a measurement mechanism for frequency spectrum occupancy rate, is characterized in that, this device guarantees to have rational confidential interval for measured frequency spectrum occupancy rate in degree of taking situation greatly, and this device comprises:

Detection module, for detection of signal strength in channel, and carries out size relatively by the signal strength detecting and thresholding, obtains comparison result sequence, wherein:

Definition random process X (t) is the random process of 2 values, and X (t)=1 represents that signal strength exceedes thresholding Li in the t moment, and X (t)=0 represents that signal strength does not exceed thresholding Li in the t moment;

X (t) is sampled, and the signal strength detecting and thresholding are carried out to size relatively, obtain comparison result sequence, this comparison result sequence is designated as random process X(n), X(n)=X _i, X(n) to obey parameter be ρ, and 0-1 distributes, X(n) probability density function be:

F (x, ρ)=ρ ^x(1-ρ) ^1-x, wherein x=0,1,0< ρ <1, n is positive integer;

Estimation module, for according to large number central-limit theorem, analyzes described comparison result sequence estimated spectral degree of taking and confidential interval thereof and relative accuracy according to Normal Distribution rule;

Wherein further comprise:

Definition stochastic variable V and W, V is launch time, and, in monitoring time T, signal strength exceedes the duration of thresholding Li continuously, and W is the time interval between transmitting, and T is much larger than the mean value of V and W;

Definition frequency spectrum occupancy rate β (L _i, T), be illustrated in monitoring time T, X (t) accounts for the ratio of total time, wherein β (L in the time of getting 1 _i, T) and be random variable of continuous type, suppose that stochastic variable W and V are separate, V+W has continuous distribution;

and within 0 to T moment, β (L _i, T) and meet progressive normal distribution, so $\mathrm{limP}[\sqrt{T}\frac{\mathrm{\&beta;}(L_i,T)-\mathrm{\&mu;}}{\mathrm{\&sigma;}}\&le;X]=\frac{1}{\sqrt{2\mathrm{\&pi;}}}{\&Integral;}_{-\&infin;}^X{e}^{-\frac{y^2}{2}}\mathrm{dy},$ Here, μ is stochastic variable β (L _i, T) mathematic expectaion;

Described estimation module specifically for:

According to large several central-limit theorems, determine:

approximate N(0,1 of obeying); Wherein,

represent sample average, i.e. the estimated value of frequency spectrum occupancy rate,

Determine: $P\{-{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\&le;\frac{n\stackrel{\&OverBar;}{X}-\mathrm{n\&rho;}}{\sqrt{\mathrm{n\&rho;}(1-\mathrm{\&rho;})}}\&le;{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}\}=1-\mathrm{\&alpha;},$ Wherein it is the normal distribution that average is 0, variance is 1

the upper boundary values at place; 1-α is confidence level;

Be equivalent to $(n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2){\mathrm{\&rho;}}^2-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2)\mathrm{\&rho;}+n{\stackrel{\&OverBar;}{X}}^2\&le;0$

If $a=n+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2,b=-(2n\stackrel{\&OverBar;}{X}+{{\mathrm{\&mu;}}_{\frac{\mathrm{\&alpha;}}{2}}}^2),c=n{\stackrel{\&OverBar;}{X}}^2,$

B ²-4ac>=0, obtains A≤ρ≤B, and wherein [A, B] is the interval estimation of ρ;

According to radical formula, obtain: $A=\frac{-b-\sqrt{b^2-4\mathrm{ac}}}{2a};B=\frac{-b+\sqrt{b^2-4\mathrm{ac}}}{2a},$

Determine: obtain relative accuracy Y:

Images