CN105682210A - SNR estimation method for PU receiver in cognitive radio system - Google Patents

Abstract

The invention belongs to the field of mobile communication, and particularly relates to the field of cognitive radio full-duplex relays in mobile communication. The SNR estimation method takes a CU as a full-duplex AF relay of a PR, a point-to-point channel between a PT and the PR becomes a three-node relay channel, receiving signals of the PR is composed of two portions: one portion is direct transmission link signals from the PT, and the other portion is relay link signals from the CU. When the amplitude gain of the CU is large enough, a second hop link signal of the relay becomes very strong, thus the signals of a direct transmission link can be neglected, and the CU and the PR has the same SNR at the moment. Therefore, the CU obtains the SNR of the PR through measuring the SNR from PT signals received by the CU. CLPT is adopted, the PT automatically maintains the SNR at the PR end as a target SNR, the SNR measured at the CU end equals the target SNR of the PR, namely a main link SNR. The SNR estimation method can effectively estimate the target SNR of the link when a G value is large.

Description

For the signal-noise ratio estimation method of primary user's receiver in cognitive radio system

Technical field

The invention belongs to moving communicating field, particularly relate to the cognitive radio full duplex relaying field in mobile communication.

Background technology

In the more than ten years in past, the frequency spectrum perception technology in cognitive radio networks is widely studied, and extremely important for frequency spectrum share. Recently, estimate the signal to noise ratio (SNR) of cross link, primary link signal power that namely cognitive user CU receives and the ratio of interfering signal power, become a problem had a great attraction. The SNR information of cross link is very important for cognitive power control room. This is because, the SNR of cross link is corresponding with the distance between master transmitter PT and cognitive user CU, and he also imply that CU is likely to the main receiver PR interference caused further: if the distance of PT-CU is very big, so CU is just likely to be located at outside the coverage of PT, seldom can have an opportunity PR is interfered; If the distance median size between PT-CU, now CU is probably close to the coverage of PT, and has more chance to interfere with PR. Therefore, transmitting power can be adjusted dynamically to control its interference level to PR based on SNR, the CU of cross link. This make when CU beyond or during close to the coverage of PT, the frequency spectrum share of Overlay and Undelay is possibly realized. But, when CU is in the coverage of PT, namely when the distance of PT-CU is smaller, descending frequency spectrum share is necessary not only for the SNR signal to noise ratio of signal (CU receive) of cross link, but also need the SNR of primary link, the namely ratio of the signal power of the master transmitter that PR termination receives and noise power. This is because the SNR of primary link provides the range information of PT-PR, this can be used to estimate that PR is likely to experienced interference. If CU and PR distance PT has similar distance, then now to be likely to distance PR close for CU, and causes serious interference. It addition, if CU and PT is bigger to the distance size differences of PT, for instance, the distance of PT-CU is only small and the distance of PT-PR is very big, then now the distance of PT and CU is likely to farther out, and namely PR can not cause very strong interference.Therefore, primary link SNR is most important for Undelay frequency spectrum share.

In existing research, few people study the SNR of primary link and estimate. This is because the SNR of primary link (i.e. the target signal to noise ratio of PT-PR) is usually obtained by PR, then pass through feedback circuit and send CU to. But in cognitive radio, such a loop is usually infeasible. Therefore for CU, estimate that the SNR of primary link is very challenging. Therefore, this patent proposes a kind of new method and removes to estimate automatically the SNR of primary link when main transceiver have employed close-loop power control CLPC. In our method, we allow CU that the signal of main transceiver carries out full duplex " amplify-forward " (AF) relaying. By measuring the signal of primary link, CU can obtain the SNR of primary link. Especially since full duplex relaying is likely to PR is interfered, control to be a default value so we devise a detector by probability of interference.

Summary of the invention

The present invention is directed to the deficiencies in the prior art, it is proposed to the method estimating automatically the SNR of primary link, and devise a detector and control in default level disturbing. In the present invention, CU can accurately estimate the SNR of primary link, and the interference that full duplex relaying is likely to result in is controlled.

In order to describe present disclosure easily, first term used in the present invention and model are introduced:

Define 1 dominant base (PrimaryTransmitter, PT): the signal transmitting terminal in main system.

Define 2 primary user PrimaryUser, PR): the signal receiving end in main system.

Define 3 cognitive user (CognitiveUser, CU): subsystem has the user of cognitive function.

Define 4 signal to noise ratios (SignalNoiseRatio, SNR): the ratio of signal power and noise power.

Define 5 close-loop power controls (ClosedLoopPowerControl, CLPC): signal transmitting terminal adjusts the transmitting power of self according to the change of signal receiving end signal to noise ratio, thus ensureing the quality of reception of receiving terminal.

Define 6 amplification forwardings (amplify-and-forward, AF): via node to the received signal, after amplifying with certain amplitude gain, relays to other receivers.

Define 7 full duplexs (Full-Duplex, FD): uplink and downlink communication is respectively adopted the identical time slot of same frequency channel and is operated.

Define 8 equivalent end-to-end channel gains (EquivalentEnd-to-EndChannelGain, ECG): between PT and PR, had the equivalent channel gain after relaying.

Define 9 times of advent poor (TimeDifferenceofArrival, TDOA): the time difference of direct transfer link signal and relay forwarding signal arrival PR end.

System used in the present invention is:

As it is shown in figure 1,

In CLPC system, PT and PR communicates in certain frequency range, and PT fixes position, the coverage position R of described PT, then set the center of circle as PT, and radius is the circle of R is circle R, PR random distribution in circle R. Meanwhile, a CU is had to want the SNR estimating the primary link at PR place to carry out the frequency spectrum share of Underlay in described round R. Distance between CU and PT is r. With(k=0,1 and 2) represents the coefficient of three inter-nodal channel, wherein, and h_kRepresent that multipath fading path loss is sparse, g_kRepresent large scale path loss coefficient. In multipath fading, described h_kObey the rayleigh distributed of unit variance. In large scale path loss, described g_kIt is expressed as g_k(dB)=-128.1-κ log₁₀L (), wherein, κ=37.6 and l >=0.035km, l represent the distance between two nodes.Consider bulk nanometer materials, then each piece interior Rayleigh fading coefficient is constant, and separate between each piece.

First, when CU keeps mourning in silence, between PT and PR, point-to-point model is introduced.

Ensured the service quality of primary user by CLPC, now, PT adjusts transmitting power automatically to ensure that PR is with constant target signal to noise ratio SNR or Signal to Interference plus Noise Ratio SINR, namelyThen obtainAnd p₀Between relational expression be:Wherein, p₀For the transmitting power of PT, (i, j) is the signal to launch of PT to x, and (i j) is unit power, namely to xFor seeking expectation computing, σ²For the PR noise power received, i represents the index of N number of sampling, and j represents the index of M block. Assuming that the noise power that CU receives is equal to σ², then the signal to noise ratio snr that CU receives is

Secondly, between PT, CU and PR, three nodal analysis methods are introduced.

When CU carries out full duplex AF relaying with gain amplifier G, the reception signal of PR end becomes:Wherein, n_cIt is the additive white Gaussian noise (AWNG) of CU end, n_pBe AWNG, the α of PR end it is sampling period T_sQuantity, β is block period T_bQuantity, total delay time T=α T_s+βT_bIt is direct signal and the TDOA of relaying signal.

Technical scheme:

Use full duplex AF relaying technique to make CU can automatically obtain the SNR of primary link, allow CU relay as the full duplex AF of PR. Then the point-to-point channel between PT and PR just becomes three node trunk channels, and the reception signal of PR is made up of two parts: a part originates from the link signal that direct transfers of PT, and another part originates from the repeated link signal of CU. When the amplitude gain of CU is sufficiently large time, as tended to just infinite, it is strong that the second hop link signal of relaying will become very, and therefore, the signal of tie link can be left in the basket. Now, CU and PR has identical SNR. Thus, CU, by measuring the SNR from dominant base signal that its own receives, obtains PR moral SNR. Adopting close-loop power control (CLPC), PT keeps the SNR of PR end to be target signal to noise ratio automatically. CU end measurement to SNR be equal to the target signal to noise ratio of PR, i.e. primary link SNR.

For the signal-noise ratio estimation method of primary user's receiver in cognitive radio system, specifically comprise the following steps that

S1, CU probability of interference according to system requirementsCalculate the decision threshold η whether carrying out relaying, wherein,For g₁≥g₀It is assumed thatFor g₁< g₀It is assumed that f₁(z) be?Probability density function under assuming, r is the distance of PT-PR, and R is the coverage of PT,The signal to noise ratio transmitting signal from PT end is monitored for CU;

In S2, downlink transmission, PT adopts adaptive modulation system to send data to PR, carries out close-loop power control, sends and keeps signal to noise ratio constant in data procedures, and CU monitors the signal to noise ratio transmitting signal from PT endAccording to decision ruleMake decisions,

If court verdict isThen illustrate that CU carries out full duplex relaying and will not worsen the channel gain of primary link, it is possible to carry out full duplex AF relaying, proceed to S3,

If court verdict isCU must keep mourning in silence, it is impossible to relays;

S3, CU are amplified with the signal of the amplitude gain G PT to receiving, and are then forwarded to PR, and now PR can receive the two paths of signals respectively from PT and CU;

S4, PR have received the signal sum of PT and CU, pass through close-loop power control so that PT adjusts the transmitting power of oneself so that the signal to noise ratio that PR receives is always target signal to noise ratio

S5, performing close-loop power control due to PT, so CU receives the signal to noise ratio of signal and is equal to the signal to noise ratio of primary link, therefore CU launches the signal to noise ratio of signal by measuring the PT oneself received, and obtains the signal to noise ratio of primary link.

The invention has the beneficial effects as follows:

The present invention can ensure that primary link is when controllable interference level, carries out full duplex AF relaying by CU, then CU can be obtained by the signal to noise ratio of primary link by measuring the signal to noise ratio of oneself. Simulation result shows, when G-value is very big, the inventive method can well estimate the target signal to noise ratio of link.

Accompanying drawing explanation

Fig. 1 is the system model figure of the present invention.

Fig. 2 is the signal to noise ratio that cognitive user receivesTwo kinds of different hypothesisWithUnder PDF curve.

Fig. 3 isThe integrating range schematic diagram of x when assuming.

Fig. 4 isThe integrating range schematic diagram of x when assuming.

Fig. 5 is under real-time situation, it is determined that obtain the amplitude peak gain G needed for target signal to noise ratio_maxSchematic diagram.

Detailed description of the invention

Below in conjunction with embodiment and accompanying drawing, describe technical scheme in detail.

As it is shown in figure 1, system has CLPC, PT and PR to communicate in certain frequency range, PT fixes position, its coverage R, then set the center of circle as PT, and the circle that radius is R is uniformly distributed in circle for circle R, PR. Meanwhile, a CU wants the SNR estimating the primary link at PR place to carry out the frequency spectrum share of downlink. Distance between CU and PT is r. With(k=0,1 and 2) represents the coefficient of three inter-nodal channel, wherein, and h_kAnd g_kRepresent multipath fading and large scale path loss coefficient respectively. In multipath fading, coefficient h_kObey the rayleigh distributed of unit variance. In large scale path loss, coefficient g_kG can be expressed as_k(dB)=-128.1-κ log₁₀L (), wherein, κ=37.6 and l >=0.035km, l represents the distance between two nodes here. Considering bulk nanometer materials, in each piece, Rayleigh fading coefficient is constant, and separate between each piece.

When CU keeps mourning in silence, introduce the point-to-point model between PT and PR.

Ensured the service quality of primary user by CLPC, now, PT adjusts transmitting power automatically to ensure PR specific target signal to noise ratio SNR or Signal to Interference plus Noise Ratio SINR, is expressed asThen obtainAnd p₀Between relational expression be:Wherein, p₀For the transmitting power of PT, (i is j) signal to launch of PT, and be unit power, namely to xWherein,For seeking expectation computing, σ²For the PR noise power received, i and j represents the index of N number of sampling and M block respectively. Assuming that the noise power that CU receives is also equal to σ², then the signal to noise ratio snr that CU receives is

Three nodal analysis methods are introduced between PT, CU and PR.

When CU carries out full duplex AF relaying with gain amplifier G, the reception signal of PR end becomes:Wherein, n_cAnd n_pBeing the additive white noise (AWNG) of CU and PR respectively, α and β is sampling period T respectively_sAnd block period T_bQuantity, and total delay time T=α T_s+βT_bIt is direct signal and the TDOA of relaying signal. In algorithmic derivation, ignore self-interference.

Primary link SNR estimates

Use full duplex AF relaying technique to realize CU and automatically obtain the SNR of primary link. Allowing CU is that PR performs full duplex AF relaying. Then, the original point-to-point channel between PT and PR has reformed into three node trunk channel, here faces, and the reception signal of PR is made up of two parts: a part originates from the direct signal of PT, and another part originates from the relaying signal of CU. When the amplitude gain of CU is sufficiently large time, such as it tends to just infinite, and it is strong that the second hop link of relaying will become very, to such an extent as to the signal of tie link can be left in the basket.At this moment, CU and PR will have identical SNR. Then, CU just by measuring the SNR of the main signal that its own receives, can obtain the SNR of PR. Have employed close-loop power control (CLPC), therefore, PT can keep the SNR of PR end to be target signal to noise ratio automatically. So, CU end measurement to SNR be equal to the target signal to noise ratio of PR, namely primary link SNR.

First considering that relaying has big amplitude gain, now, PR will treat through and relaying signal in a different manner. Then, when amplitude gain tends to just infinite, obtain and make the CU amplitude gain obtained required for best track performance.

When CU performs full duplex AF relaying, main signal, by different paths, then arrives PR with the different time. Considering system bandwidth and signal handling capacity, PR is merely able within a specific time period to receive signal, and this is called maximum available TDOA, is expressed as T_m. If the time delay difference of direct signal and relaying signal is less than T_m, i.e. τ < T_m, this is little time delay situation, and now, PR regards direct signal and relaying signal as the signal of needs as. But, if TDOA is equal to or more than T_m, i.e. τ > T_m, this is long time delay situation, now PR only using two paths of signals Zhong mono-road as the signal needed, and interference is regarded as in an other road.

The method of the present invention that derives under little time delay and long time delay situation separately below.

Little time delay situation:

PR regards desired signal as direct signal and relaying signal, it is possible to obtained the SNR of PR end by following formula:Wherein, g_e=(g₀+G²g₁g₂)/(G²g₂+ 1) it is of equal value end-to-end channel gain (ECG). When amplitude gain G is sufficiently large, it is possible to obtain:Then basisObtain the SNR of primary link, wherein,It it is the transmitting power of PT when there being relaying.

Long time delay situation:

PR can select signal intensity compared with great Na road signal as desired signal automatically, and signal more weak for an other road is regarded as interference signal. Owing to CU employs big amplitude gain when relaying, therefore the intensity of relaying signal is more than the intensity of direct signal. Following formula can be passed through, obtain the average SINR of PR end:Wherein, ECCG becomes g_e=G²g₁g₂/(p₀g₀/σ²+G²g₂+ 1), when amplitude gain is sufficiently large, obtain:Then according to formulaObtain the SNR of primary link.

From deriving it can be seen that when CU relays using big amplitude gain as full duplex AF, it can estimate the SNR of primary link above. But, on the other hand, when CU relays as AF full duplex, it is possible to PR can be interfered. This is because, relay processes is actually channel gain g original between PT-PR₀Become ECGg_e=g₁If: g₁< g₀, it means that PR is brought interference by relaying, is degrading primary link channel; If g₁≥g₀, then relay and PR do not interfered. Therefore, CU is required to identify g₁≥g₀Situation so that PR will not be interfered by CU.

Interference management

DefinitionWithRepresent g respectively₁< g₀And g₁≥g₀Two kinds of hypothesis. Based on this, a kind of detector of design goes to identify situation g₁≥g₀, namely trunk channel first jumps the gain gain more than direct channel. This make CU can when meet preset the given probability of interference to PR, use the method for estimation of the present invention.

Find test statistics

Owing to CU and the PR signal received both is from a common transmitter PT, so the target signal to noise ratio of PRAnd the signal to noise ratio that CU measurement obtainsAll by corresponding channel gain g₀And g₁Determined, it is possible to obtainAs can be seen from the above equation, g₁And g₀Relation, corresponding toWithRelation.So, CU can pass through to compare the signal to noise ratio that its measurement obtainsAnd the target signal to noise ratio of PRJust can detect whether present case meets g₁≥g₀. Therefore, the measurement signal to noise ratio of CU end is selectedAs test statistics.

The derivation of PDF expression formula:

According toThe test statistics obtained is:Employing dB is unit, willIt is rewritten as

OrderThen z=x+y.

Assuming that the target signal to noise ratio SNR of PR, namelyDescribedIt is a stochastic variable, obeys γ_LAnd γ_HBetween be uniformly distributed. Then corresponding probability density function is f (x)=1/ (γ_H-γ_L), wherein, x ∈ [γ_L,γ_H], owing to PR is evenly distributed in the coverage of PT, the probability density function obtaining y is as follows:Wherein,Correspondence and hypothesis Correspondence and hypothesis

Owing to x, y the two stochastic variable is independent from, then according to f (z)=∫ f (x) f (z-x) dx, obtain the PDF of test statistics z.

AssumeY=z-x and κ lg (ε/r)≤y < 0 can be obtained, i.e. z≤x < z-κ lg (ε/r), represent the integrating range of variable x.

If the distance r of PT-CU is smaller, namelyThen

Wherein,

$\mathrm{\&phi;}=r^2\&CenterDot;{10}^{\frac{2z}{\mathrm{\&kappa;}}}/\left(\right({\mathrm{\&gamma;}}_H-{\mathrm{\&gamma;}}_L\left)\right(r^2-{\mathrm{\&epsiv;}}^2\left)\right).$

If the distance between PT-CU is median size, namely

$f_1\left(z\right)=\left\{\begin{array}{cc}({10}^{-\frac{2{\mathrm{\&gamma;}}_L}{\mathrm{\&kappa;}}}-{10}^{-\frac{2(z-{\mathrm{\&gamma;}}_L+{\mathrm{\&gamma;}}_H)}{\mathrm{\&kappa;}}})\mathrm{\&phi;},& 2{\mathrm{\&gamma;}}_L-{\mathrm{\&gamma;}}_H\&le;z<{\mathrm{\&gamma;}}_L,\\ ({10}^{-\frac{2z}{\mathrm{\&kappa;}}}-{10}^{-\frac{2{\mathrm{\&gamma;}}_H}{\mathrm{\&kappa;}}})\mathrm{\&phi;},& {\mathrm{\&gamma;}}_L\&le;z<{\mathrm{\&gamma;}}_H,\end{array}\right.$

If the distance between PT-CU is relatively larger, namelyThen

$f_1\left(z\right)=\left\{\begin{array}{cc}\left({10}^{-\frac{2{\mathrm{\&gamma;}}_L}{\mathrm{\&kappa;}}}-{10}^{-\frac{2\left(z-\mathrm{\&kappa;}\lg \left(\frac{\mathrm{\&epsiv;}}{r}\right)\right)}{\mathrm{\&kappa;}}}\right)\mathrm{\&phi;},& \mathrm{\&kappa;}\lg \left(\frac{\mathrm{\&epsiv;}}{r}\right)+{\mathrm{\&gamma;}}_L\&le;z<\mathrm{\&kappa;}\lg \left(\frac{\mathrm{\&epsiv;}}{r}\right)+{\mathrm{\&gamma;}}_H,\\ \left({10}^{-\frac{2{\mathrm{\&gamma;}}_L}{\mathrm{\&kappa;}}}-{10}^{-\frac{2{\mathrm{\&gamma;}}_H}{\mathrm{\&kappa;}}}\right)\mathrm{\&phi;},& \mathrm{\&kappa;}\lg \left(\frac{\mathrm{\&epsiv;}}{r}\right)+{\mathrm{\&gamma;}}_H\&le;z<{\mathrm{\&gamma;}}_L,\\ \left({10}^{-\frac{2z}{\mathrm{\&kappa;}}}-{10}^{-\frac{2{\mathrm{\&gamma;}}_H}{\mathrm{\&kappa;}}}\right)\mathrm{\&phi;},& {\mathrm{\&gamma;}}_L\&le;z\&le;{\mathrm{\&gamma;}}_H,\end{array}\right.$

Assume

And hypothesisSituation similar, due to x integration interval meetThe expression formula of following PDF can be obtained:

If the distance between PT-CU is medium, namelyThen

If the distance between PT-CU is relatively larger, namelyThen

Find decision threshold:

As shown in Figure 2, test statisticsTwo kinds of hypothesisWithUnder PDF curve, wherein PR is evenly distributed in the coverage of PT, the distance r=0.2km between PT-PR. Especially, it is contemplated thatWithPrior information. It will be seen that the PDF curve that both is assumed is different from figure. So, CU is distinguished by following decision ruleWithTwo kinds of situations:Wherein, η is thresholding. That is, the signal to noise ratio that if CU measurement is arrivedMore than threshold value η, then CU will be allowed to carry out full duplex relaying. Otherwise, CU must keep mourning in silence, to avoid PT is interfered.

In order to meet specific probability of interference P_I, threshold value η can according to formula

Obtained by numerical computations in simulations.

Simulation result

Adopting the system model identical with accompanying drawing 1, the covering radius of PT is R=0.5km, and the distance of PT-CU is rkm. In simulations, target signal to noise ratioAt γ_L=5dB and γ_HIt is uniformly distributed between=30dB, probability of interference P_I=5%, noise power is-114dBm, and the quantity of block is M=200, and in each piece, the quantity of sampling is N=100, and the number of times of Monte Carlo Experiment is 10³. For wireless channel, it is contemplated that path loss, shade and multipath fading, wherein shadow factor obeys standard deviation is the logarithm normal distribution of 4. Full duplex AF is relayed, it is contemplated that the self-interference of imperfections suppresses. Improve noise gate to 2dB, self-interference is tapered to noise level.

As it is shown in figure 5, the normalization error of the method for the present invention, wherein the distance of PT-CU is contemplated. Normalized error is defined asFrom figure 5 it can be seen that when with and without shade, estimation difference is similar, and they increase along with amplitude gain G and decline. Especially, if G >=80dB, as r=0.2km and r=0.45km, estimation difference has reached 0.025. Further, when considering the self-interference of imperfections, estimation difference has deteriorated into about 0.05.But, minimum estimation difference is still respectively smaller than 0.05 and 0.08 when r=0.2km and r=0.45km.

Claims (1)

1. for the signal-noise ratio estimation method of primary user's receiver in cognitive radio system, it is characterised in that comprise the steps:

S1, CU probability of interference according to system requirementsCalculate the decision threshold η whether carrying out relaying, wherein,For g₁≥g₀It is assumed thatFor g₁< g₀It is assumed that f₁(z) be?Probability density function under assuming, r is the distance of PT-PR, and R is the coverage of PT,The signal to noise ratio transmitting signal from PT end is monitored for CU;

In S2, downlink transmission, PT adopts adaptive modulation system to send data to PR, carries out close-loop power control, sends and keeps signal to noise ratio constant in data procedures, and CU monitors the signal to noise ratio transmitting signal from PT endAccording to decision ruleMake decisions,

If court verdict isThen illustrate that CU carries out full duplex relaying and will not worsen the channel gain of primary link, it is possible to carry out full duplex AF relaying, proceed to S3,

If court verdict isCU must keep mourning in silence, it is impossible to relays;

S3, CU are amplified with the signal of the amplitude gain G PT to receiving, and are then forwarded to PR, and now PR can receive the two paths of signals respectively from PT and CU;

S4, PR have received the signal sum of PT and CU, pass through close-loop power control so that PT adjusts the transmitting power of oneself so that the signal to noise ratio that PR receives is always target signal to noise ratio

S5, performing close-loop power control due to PT, so CU receives the signal to noise ratio of signal and is equal to the signal to noise ratio of primary link, therefore CU launches the signal to noise ratio of signal by measuring the PT oneself received, and obtains the signal to noise ratio of primary link.