CN105141385B - Multiband cooperative cognitive frequency spectrum sensing method - Google Patents

Abstract

The present invention relates to multiband cooperative cognitive frequency spectrum sensing method, itself signal to noise ratio, confidence level, business demand and frequency spectrum detecting result to multiple primary users are sent to frequency spectrum perception fusion center by each user respectively, calculated by frequency spectrum perception fusion center, screen primary election cooperation time user, then distributed number of frequency bands and give primary election cooperation secondary user；The quotient between the signal to noise ratio of all primary election time user's signal to noise ratio root-mean-square value and each primary election time user is calculated, according to the relation between each quotient and signal to noise ratio predetermined threshold value, the final election time user for participating in cooperation is selected, and adjust the detection probability of final election time user；With reference to the average detected probability after the signal to noise ratio of each final election time user and adjustment, the step of returning to selection final election time user, obtain the final whole choosing time user for participating in cooperation, and the global detection probability cooperated using the OR criterions of weighting is the final detection result of frequency spectrum perception fusion center, it is to avoid baneful influence of the secondary user of low signal-to-noise ratio or poor performance to whole detection performance.

Description

Multiband cooperative cognitive frequency spectrum sensing method

Technical field

The present invention relates to frequency spectrum detection field, more particularly to a kind of multiband cooperative cognitive frequency spectrum sensing method.

Background technology

It is the new of mark with LTE, Wi-Fi, satellite communication and communication for coordination etc. with continuing to develop for wireless communication technology Emerging technology is emerged in large numbers in succession, is emerged in an endless stream.These communication technologys propose higher demand to radio spectrum resources, so as to make frequency spectrum What resource became tends to be nervous, and cognitive radio technology (Cognitive Radio, CR) arises at the historic moment in this context.

The Basic Ways of cognitive radio are that user's (or perceiving user, cognitive user) secondary first uses frequency spectrum perception The frequency spectrum resource of mandate in surrounding environment is persistently detected；Then ensureing that primary user (also known as authorized user) can be excellent First take this section of frequency spectrum and under conditions of transmission performance is barely affected, secondary user is adaptively adjusted transceiver, and will Transceiver is adjusted to be communicated to idle frequency spectrum.When secondary user perceives (or detection) to when having primary user's signal to occur, Secondary user then will quickly vacate channel and be used for primary user, and then avoid the proper communication to primary user from disturbing, so as to carry For frequency spectrum resource utilization rate.

In order to reduce the factors such as multipath fading in actual environment, shadow effect and incorrect noise to detection performance Adverse effect, the cooperative frequency spectrum sensing method based on multiple users constantly proposed.By by the detection of each user As a result it is sent to frequency spectrum perception fusion center to be merged, to reach the purpose for perceiving frequency spectrum.

However, existing cooperative frequency spectrum sensing method majority is perceived just for one-segment, in order to improve frequency spectrum profit With rate, the collaborative sensing method for multiband turns into new study hotspot.But, in the collaborative sensing of multiband, due to The detection performance of each user, signal to noise ratio are entirely not excellent, when all secondary users both participate in the association to multiband When perceiving, the secondary user of low signal-to-noise ratio or detection poor performance can make a very bad impression to overall detection performance.

The content of the invention

The technical problems to be solved by the invention are to provide one kind for above-mentioned prior art to avoid with low noise Than or detection poor performance the multiband cooperative cognitive frequency spectrum sensing method that is made a very bad impression to whole detection performance of secondary user.

The present invention solve the technical scheme that is used of above-mentioned technical problem for：Multiband cooperative cognitive frequency spectrum sensing method, Characterized in that, in turn including the following steps：

(1) it is set in cognition network, the quantity of primary user is N₁, the quantity of secondary user is N₂, frequency spectrum perception fusion center Quantity be 1, primary user separately takes the respective frequency range in frequency spectrum；N₂Individual user separately obtains itself letter Make an uproar and compare SNR_jAnd to N₁Individual primary user takes the frequency spectrum detecting result of frequency range, and respectively by the signal to noise ratio snr of acquisition_j, frequency spectrum inspection Survey result, detection confidence level P_jWith business demand R_jSend to frequency spectrum perception fusion center, wherein,

Primary user is designated as PU_i, secondary user is designated as CR_j, frequency spectrum perception fusion center is designated as FC, business demand R_j∈[0,N₁], P_j∈ [0,1], detects confidence levelP_d,jiFor secondary user CR_jTo primary user PU_iDetection probability；1≤i≤ N₁, 1≤j≤N₂, N₁>=2, N₂≥2；

(2) frequency spectrum perception fusion center FC receives each user CR_jThe signal to noise ratio snr sent_j, frequency spectrum detecting result, inspection Survey confidence level P_jWith business demand R_j, and judge the signal to noise ratio snr of time user_jMore than default signal to noise ratio screening value SNR_choseWhen, Select corresponding user of this signal to noise ratio to participate in the primary election time user of cooperative detection, and remember that primary election time user is CR'_t, and hold Row step (3), otherwise, the frequency spectrum detecting result corresponding to secondary user of the selection with highest signal to noise ratio is in frequency spectrum perception fusion Heart FC final detection result；Wherein,

Primary election time number of users is N'₂, primary election time user CR'_tCorresponding signal to noise ratio is SNR'_t, detection confidence level be P'_t、 Business demand is R'_t, 1≤t≤N'₂≤N₂；The frequency spectrum detecting result quantity that frequency spectrum perception fusion center FC is received is N₁×N₂ It is individual；

(3) frequency spectrum perception fusion center FC is according to primary election time user CR'_tConfidence level P'_t, business demand R'_t, to primary election Secondary user CR'_tDistribution needs the number of frequency bands C detected_t：

(3-1) is according to each primary election time user CR'_tConfidence level P'_t, respectively to each primary election time user CR'_tConfidence Spend P'_tIt is normalized, obtains each primary election time user CR'_tNormalization confidence value

(3-2) each primary election time user CR' according to obtained by step (3-1)_tCorresponding normalization confidence valueCalculate frequency Spectrum perceives fusion center FC distribution primary election time users CR'_tNeed the number of frequency bands C of detection_t：

(4) frequency spectrum perception fusion center FC is according to the primary election time user CR' for participating in cooperative detection_tSignal to noise ratio snr '_t, meter Calculate the signal to noise ratio root-mean-square value of all primary election time userAnd make signal to noise ratio snr '_t=γ_t, wherein, signal to noise ratio root-mean-square value Be calculated as follows：

(5) frequency spectrum perception fusion center FC calculates the signal to noise ratio root-mean-square value of all primary election time user successively respectivelyWith it is each Primary election time user CR'_tSignal to noise ratio snr '_tBetween quotient η_t, wherein,

(6) frequency spectrum perception fusion center FC is calculated, is obtained signal to noise ratio predetermined threshold value λ and signal to noise ratio optimal threshold λ_optimal, And respectively according to each signal to noise ratio quotient η_tWith the magnitude relationship between signal to noise ratio predetermined threshold value λ, the final election time for participating in cooperation is selected User CR "_k, final election time user CR "_kSignal to noise ratio be SNR "_k, wherein,

(6-1) frequency spectrum perception fusion center FC is according to the N' of reception₂Individual primary election is from the corresponding signal to noise ratio set of user {SNR'_t, primary election is obtained from user's signal to noise ratio set { SNR'_tIn signal to noise ratio maximum, remember the signal to noise ratio maximum be SNR' max；

(6-2) using the signal to noise ratio maximum SNR'max of acquisition as reference, and by N'₂Individual primary election is from user CR'_tSignal to noise ratio SNR'_tMake business's processing with signal to noise ratio maximum SNR'max respectively, calculating obtain each primary election from user's signal to noise ratio snr '_tCorresponding Initial threshold λ_t, wherein,

λ_t=| SNR'_t/SNR'_max|, t=1,2 ..., N'₂, N'₂≤N₂；

(6-3) is according to each primary election from user CR'_tNormalization confidence valueWith signal to noise ratio quotient η_t, calculate each primary election From user CR'_tJoint screening parameter value ξ_t, and according to joint screening parameter value ξ_t, the final election for participating in cooperation is chosen from user CR”_k, wherein, final election is from user CR "_kQuantity be M,T=1,2 ..., N'₂, k=1,2 ..., M, M≤N'₂：

If joint screening parameter value ξ_tPositioned at default value interval range [ξ_a,ξ_b] in, i.e. ξ_a≤ξ_t≤ξ_bWhen, then choosing should Joint screening parameter value ξ_tCorresponding primary election from user be final election from user, and participate in cooperative detection；Otherwise, the primary election is from user Not choose；

The signal to noise ratio predetermined threshold value λ of (6-4) in step (6-3), obtains M final election from user CR "_kIt is accurate in OR respectively Then with the cooperative detection performance curve under AND criterions, wherein,

OR criterions：

AND criterions：K=1,2 ..., M, M≤N'₂≤N₂；

Wherein, P_d,kIt is k-th of final election from user CR "_kAverage detected probability, P_fa,kIt is k-th of final election from user CR "_k Average false-alarm probability；P_d,ksIt is final election from user CR "_kTo the detection probability of its allocated s-th of frequency range, P_fa,ksFor final election from User CR "_kTo the false-alarm probability of its allocated s-th of frequency range；Q_dFor the global detection probability after cooperative detection, Q_faFor cooperation inspection Global false-alarm probability after survey；ω_kRepresent signal to noise ratio CR "_kWeight coefficient, SNR "_kIt is k-th of final election from user CR "_kLetter Make an uproar and compare, SNR "_maxRepresent signal to noise ratio maximum of the M final election from user, SNR "_minRepresent M final election from the signal to noise ratio of user most Small value；

(6-5) is respectively obtained accurate in OR criterions and AND according to the cooperative detection performance curve under OR criterions and AND criterions Maximum detection probability Q under then_{(OR, d)-max}、Q_{(AND, d)-max}, obtain Q_{(OR, d)-max}And Q_{(AND, d)-max}Maximum Q_d-max, and with this Optimum detection performance value Q_d-maxCorresponding signal to noise ratio predetermined threshold value is signal to noise ratio optimal threshold, and note signal to noise ratio optimal threshold is λ_optimal；Wherein, Q_d-max=max (Q_{(OR, d)-max}, Q_{(AND, d)-max})；

(7) according to the signal to noise ratio optimal threshold λ of acquisition_optimal, obtain signal to noise ratio optimal threshold λ_optimalCorresponding final election From user CR ", the final election is obtained from user CR " Dynamic gene α and other M-1 final election from user CR "_kDynamic gene α_k, and respectively according to Dynamic gene α, α_kCorrespondence adjusts final election from user CR ", CR "_kAverage false-alarm probability, final election is from user Average false-alarm probability after CR " adjustment is designated as P_fa, final election is from user CR "_kAverage false-alarm probability after adjustment is designated as P_fa,k；Its In,

P_fa,k=α_k·P_fa, k=1,2 ..., M-1；

Wherein, α_kIt is final election from user CR "_kDynamic gene, for according to final election from user CR "_kThe signal to noise ratio of itself SNR”_kRealize the adjustment to its average false-alarm probability size；SNR”_kIt is k-th of final election from user CR "_kSignal to noise ratio；

(8) according to the M final election obtained in step (7) from the Dynamic gene α of user_kAnd the average void after correspondence adjustment Alarm probability P_fa,k, final election is calculated from user CR "_kJudgement threshold values λ " after adjustment_kWith average detection probability P_d,k, wherein,

Wherein,K=1,2 ..., M, M≤N'₂；N is sampling number；

(9) according to signal to noise ratio snr of the M final election from user in step (8) "_kAnd the average detected after obtained adjustment Probability P_d,k, return to step (6) selects in M final election from user again, and obtain participating in cooperation T selects from user eventually CR”'_t, and using the global detection probability after the OR criterions cooperation of weighting as frequency spectrum perception fusion center FC final detection result, Wherein 1≤t≤T≤M≤N'₂。

Further, the OR criterions of weighting are as follows in the step (9)：

Wherein, P'_d,tsTo select eventually from user CR " '_tTo the detection probability of its allocated s-th of frequency range, P'_fa,tsFor whole choosing From user CR " '_tTo the false-alarm probability of its allocated s-th of frequency range；P'_d,tSelected for the end that t-th reselects from user CR " '_t Average detected probability, P'_fa,tSelected for the end that t-th reselects from user CR " '_tAverage false-alarm probability；Q'_dFor cooperation inspection Global detection probability after survey, Q'_faFor the global false-alarm probability after cooperative detection；M' is the whole number selected from user reselected Mesh；ω'_tIt is the whole choosing that reselects from user CR " '_tWeight coefficient.

Compared with prior art, the advantage of the invention is that：Each user is respectively by itself signal to noise ratio, confidence level, business Demand and frequency spectrum detecting result to multiple primary users are sent to frequency spectrum perception fusion center, by frequency spectrum perception fusion center meter The secondary user of primary election cooperation time user, deletion detection poor performance and low signal-to-noise ratio are calculated, screen, then distribution number of frequency bands is to primary election Cooperation time user；Calculate the quotient between the signal to noise ratio of all primary election time user's signal to noise ratio root-mean-square value and each primary election time user, root According to the relation between each quotient and signal to noise ratio predetermined threshold value, the final election time user for participating in cooperation is selected, and adjust final election time user's Detection probability；With reference to the average detected probability after the signal to noise ratio of each final election time user and adjustment, return to selection final election time and use The step of family, the whole choosing time user of final participation cooperation is obtained, and using the global detection probability of the OR criterions cooperation of weighting as frequency Spectrum perceives the final detection result of fusion center, it is to avoid evil of the secondary user of low signal-to-noise ratio or poor performance to whole detection performance Bad influence.

Brief description of the drawings

Fig. 1 is the cognition network structural representation in the embodiment of the present invention；

Fig. 2 is the multiband cooperative cognitive frequency spectrum sensing method schematic flow sheet of the embodiment of the present invention；

Multiband cooperative cognitive frequency spectrum sensing methods and tradition AND criterion collaboration frequency spectrum sense of the Fig. 3 for the embodiment of the present invention The simulation result schematic diagram known.

Embodiment

The present invention is described in further detail below in conjunction with accompanying drawing embodiment.

As shown in figure 1, being set in cognition network, the quantity of primary user is N₁, the quantity of secondary user is N₂, frequency spectrum perception The quantity of fusion center is 1, wherein, primary user is designated as PU_i, secondary user is designated as CR_j, frequency spectrum perception fusion center is designated as FC；1≤i ≤N₁, 1≤j≤N₂, N₁>=2, N₂≥2。

Below in conjunction with Fig. 1 and Fig. 2, multiband cooperative cognitive frequency spectrum sensing method in the present embodiment is made an explanation.Wherein, The multiband cooperative cognitive frequency spectrum sensing method, in turn includes the following steps：

(1) it is set in cognition network, the quantity of primary user is N₁, the quantity of secondary user is N₂, frequency spectrum perception fusion center Quantity be 1, primary user separately takes the respective frequency range in frequency spectrum；N₂Individual user separately obtains itself letter Make an uproar and compare SNR_jAnd to N₁Individual primary user takes the frequency spectrum detecting result of frequency range, and respectively by the signal to noise ratio snr of acquisition_j, frequency spectrum inspection Survey result, detection confidence level P_jWith business demand R_jSend to frequency spectrum perception fusion center, wherein,

Primary user is designated as PU_i, secondary user is designated as CR_j, frequency spectrum perception fusion center is designated as FC, business demand R_j∈[0,N₁], P_j∈ [0,1], detects confidence levelP_d,jiFor secondary user CR_jTo primary user PU_iDetection probability；1≤i≤ N₁, 1≤j≤N₂, N₁>=2, N₂≥2；

(2) frequency spectrum perception fusion center FC receives each user CR_jThe signal to noise ratio snr sent_j, frequency spectrum detecting result, inspection Survey confidence level P_jWith business demand R_j, and judge the signal to noise ratio snr of time user_jMore than default signal to noise ratio screening value SNR_choseWhen, Select corresponding user of this signal to noise ratio to participate in the primary election time user of cooperative detection, and remember that primary election time user is CR'_t, and hold Row step (3), otherwise, the frequency spectrum detecting result corresponding to secondary user of the selection with highest signal to noise ratio is in frequency spectrum perception fusion Heart FC final detection result；Wherein,

Primary election time number of users is N'₂, primary election time user CR'_tCorresponding signal to noise ratio is SNR'_t, detection confidence level be P'_t、 Business demand is R'_t, 1≤t≤N'₂≤N₂；The frequency spectrum detecting result quantity that frequency spectrum perception fusion center FC is received is N₁×N₂ It is individual；

(3) frequency spectrum perception fusion center FC is according to primary election time user CR'_tConfidence level P'_t, business demand R'_t, to primary election Secondary user CR'_tDistribution needs the number of frequency bands C detected_t：

(3-1) is according to each primary election time user CR'_tConfidence level P'_t, respectively to each primary election time user CR'_tConfidence Spend P'_tIt is normalized, obtains each primary election time user CR'_tNormalization confidence value

(3-2) each primary election time user CR' according to obtained by step (3-1)_tCorresponding normalization confidence valueCalculate frequency Spectrum perceives fusion center FC distribution primary election time users CR'_tNeed the number of frequency bands C of detection_t：

(4) frequency spectrum perception fusion center FC is according to the primary election time user CR' for participating in cooperative detection_tSignal to noise ratio snr '_t, meter Calculate the signal to noise ratio root-mean-square value of all primary election time userAnd make signal to noise ratio snr '_t=γ_t, wherein, signal to noise ratio root-mean-square value Be calculated as follows：

(5) frequency spectrum perception fusion center FC calculates the signal to noise ratio root-mean-square value of all primary election time user successively respectivelyWith it is each Primary election time user CR'_tSignal to noise ratio snr '_tBetween quotient η_t, wherein,

(6) frequency spectrum perception fusion center FC is calculated, is obtained signal to noise ratio predetermined threshold value λ and signal to noise ratio optimal threshold λ_optimal, And respectively according to each signal to noise ratio quotient η_tWith the magnitude relationship between signal to noise ratio predetermined threshold value λ, the final election time for participating in cooperation is selected User CR "_k, final election time user CR "_kSignal to noise ratio be SNR "_k, wherein,

(6-1) frequency spectrum perception fusion center FC is according to the N' of reception₂Individual primary election is from the corresponding signal to noise ratio set of user {SNR'_t, primary election is obtained from user's signal to noise ratio set { SNR'_tIn signal to noise ratio maximum, remember the signal to noise ratio maximum be SNR' max；

(6-2) using the signal to noise ratio maximum SNR'max of acquisition as reference, and by N'₂Individual primary election is from user CR'_tSignal to noise ratio SNR'_tMake business's processing with signal to noise ratio maximum SNR'max respectively, calculating obtain each primary election from user's signal to noise ratio snr '_tCorresponding Initial threshold λ_t, wherein,

λ_t=| SNR'_t/SNR'_max|, t=1,2 ..., N'₂, N'₂≤N₂；

(6-3) is according to each primary election from user CR'_tNormalization confidence valueWith signal to noise ratio quotient η_t, calculate each primary election From user CR'_tJoint screening parameter value ξ_t, and according to joint screening parameter value ξ_t, the final election for participating in cooperation is chosen from user CR”_k, wherein, final election is from user CR "_kQuantity be M,T=1,2 ..., N'₂, k=1,2 ..., M, M≤N'₂：

If joint screening parameter value ξ_tPositioned at default value interval range [ξ_a,ξ_b] in, i.e. ξ_a≤ξ_t≤ξ_bWhen, then choosing should Joint screening parameter value ξ_tCorresponding primary election from user be final election from user, and participate in cooperative detection；Otherwise, the primary election is from user Not choose；

The signal to noise ratio predetermined threshold value λ of (6-4) in step (6-3), obtains M final election from user CR "_kIt is accurate in OR respectively Then with the cooperative detection performance curve under AND criterions, wherein,

OR criterions：

AND criterions：K=1,2 ..., M, M≤N'₂≤N₂；

Wherein, P_d,kIt is k-th of final election from user CR "_kAverage detected probability, P_fa,kIt is k-th of final election from user CR "_k Average false-alarm probability；P_d,ksIt is final election from user CR "_kTo the detection probability of its allocated s-th of frequency range, P_fa,ksFor final election from User CR "_kTo the false-alarm probability of its allocated s-th of frequency range；Q_dFor the global detection probability after cooperative detection, Q_faFor cooperation inspection Global false-alarm probability after survey；ω_kRepresent signal to noise ratio CR "_kWeight coefficient, SNR "_kIt is k-th of final election from user CR "_kLetter Make an uproar and compare, SNR "_maxRepresent signal to noise ratio maximum of the M final election from user, SNR "_minRepresent M final election from the signal to noise ratio of user most Small value；

(6-5) is respectively obtained accurate in OR criterions and AND according to the cooperative detection performance curve under OR criterions and AND criterions Maximum detection probability Q under then_{(OR, d)-max}、Q_{(AND, d)-max}, obtain Q_{(OR, d)-max}And Q_{(AND, d)-max}Maximum Q_d-max, and with this Optimum detection performance value Q_d-maxCorresponding signal to noise ratio predetermined threshold value is signal to noise ratio optimal threshold, and note signal to noise ratio optimal threshold is λ_optimal；Wherein, Q_d-max=max (Q_{(OR, d)-max}, Q_{(AND, d)-max})；

(7) according to the signal to noise ratio optimal threshold λ of acquisition_optimal, obtain signal to noise ratio optimal threshold λ_optimalCorresponding final election From user CR ", the final election is obtained from user CR " Dynamic gene α and other M-1 final election from user CR "_kDynamic gene α_k, and respectively according to Dynamic gene α, α_kCorrespondence adjusts final election from user CR ", CR "_kAverage false-alarm probability, final election is from user Average false-alarm probability after CR " adjustment is designated as P_fa, final election is from user CR "_kAverage false-alarm probability after adjustment is designated as P_fa,k；Its In,

P_fa,k=α_k·P_fa, k=1,2 ..., M-1；

Wherein, α_kIt is final election from user CR "_kDynamic gene, for according to final election from user CR "_kThe signal to noise ratio of itself SNR”_kRealize the adjustment to its average false-alarm probability size；SNR”_kIt is k-th of final election from user CR "_kSignal to noise ratio；

(8) according to the M final election obtained in step (7) from the Dynamic gene α of user_kAnd the average void after correspondence adjustment Alarm probability P_fa,k, final election is calculated from user CR "_kJudgement threshold values λ " after adjustment_kWith average detection probability P_d,k, wherein,

Wherein,K=1,2 ..., M, M≤N'₂；N is sampling number；

(9) according to signal to noise ratio snr of the M final election from user in step (8) "_kAnd the average detected after obtained adjustment Probability P_d,k, return to step (6) selects in M final election from user again, and obtain participating in cooperation T selects from user eventually CR”'_t, and using the global detection probability after the OR criterions cooperation of weighting as frequency spectrum perception fusion center FC final detection result, Wherein, the OR criterions of weighting are as follows：

Wherein, P'_d,tsTo select eventually from user CR " '_tTo the detection probability of its allocated s-th of frequency range, P'_fa,tsFor whole choosing From user CR " '_tTo the false-alarm probability of its allocated s-th of frequency range；P'_d,tSelected for the end that t-th reselects from user CR " '_t Average detected probability, P'_fa,tSelected for the end that t-th reselects from user CR " '_tAverage false-alarm probability；Q'_dFor cooperation inspection Global detection probability after survey, Q'_faFor the global false-alarm probability after cooperative detection；M' is the whole number selected from user reselected Mesh；ω'_tIt is the whole choosing that reselects from user CR " '_tWeight coefficient, 1≤t≤T≤M≤N'₂。

Fig. 3 gives the simulation result schematic diagram of multiband cooperative cognitive frequency spectrum sensing method in the present invention, and simultaneously right Traditional AND cooperative detections method is emulated.Wherein, in traditional AND cooperative detection methods, all times with joining per family Detected with collaborative spectrum sensing.Simulated conditions are as follows：

It is set in cognitive radio networks, primary user PU quantity N₁Respectively 2,3,5 and 7, from user CR quantity N₂ Gradually 11 are increased to from 2；From the signal to noise ratio of user be respectively -5dB, -8dB, -10dB, -13dB, -16dB, -17dB, - 19dB, -23dB, -25dB and -27dB, from per family using energy measuring.

As seen from Figure 3, in primary user's quantity and under conditions of number of users is certain, the global detection in the present invention Probability is greater than the detection probability of traditional AND cooperative detections, and this cooperative frequency spectrum sensing method shown in the present invention has more preferable Detection performance；It is also possible to find, under conditions of cooperative detection method and primary user's quantity are certain, with from user Global detection probability after the increase of quantity, cooperative detection gradually increases；In cooperative detection method and certain from number of users Under the conditions of, with the increase of primary user's quantity, the global detection probability after cooperative detection is then gradually reduced.Understand, it is relatively conventional AND cooperative detection methods, the multiband Cognitive-Cooperation frequency spectrum sensing method in the embodiment of the present invention is because avoiding low signal-to-noise ratio Or the baneful influence of the secondary user of detection poor performance, overall cooperative detection performance is greatly improved, with more preferable inspection Survey performance.

Claims (1)

1. multiband cooperative cognitive frequency spectrum sensing method, it is characterised in that in turn include the following steps：

(1) it is set in cognition network, the quantity of primary user is N₁, the quantity of secondary user is N₂, the number of frequency spectrum perception fusion center Measure as 1, primary user separately takes the respective frequency range in frequency spectrum；N₂Individual time user separately obtains itself signal to noise ratio SNR_jAnd to N₁Individual primary user takes the frequency spectrum detecting result of frequency range, and respectively by the signal to noise ratio snr of acquisition_j, frequency spectrum detection knot Really, detection confidence level P_jWith business demand R_jSend to frequency spectrum perception fusion center, wherein,

Primary user is designated as PU_i, secondary user is designated as CR_j, frequency spectrum perception fusion center is designated as FC, business demand R_j∈[0,N₁], P_j∈ [0,1], detects confidence levelP_d,jiFor secondary user CR_jTo primary user PU_iDetection probability；1≤i≤N₁, 1≤ j≤N₂, N₁>=2, N₂≥2；

(2) frequency spectrum perception fusion center FC receives each user CR_jThe signal to noise ratio snr sent_j, frequency spectrum detecting result, detection put Reliability P_jWith business demand R_j, and judge the signal to noise ratio snr of time user_jMore than default signal to noise ratio screening value SNR_choseWhen, selection Corresponding user of this signal to noise ratio remembers that primary election time user is CR' to participate in the primary election time user of cooperative detection_t, and perform step Suddenly (3), otherwise, the frequency spectrum detecting result corresponding to secondary user of the selection with highest signal to noise ratio is frequency spectrum perception fusion center FC Final detection result；Wherein,

Primary election time number of users is N'₂, primary election time user CR'_tCorresponding signal to noise ratio is SNR'_t, detection confidence level be P'_t, business Demand is R'_t, 1≤t≤N'₂≤N₂；The frequency spectrum detecting result quantity that frequency spectrum perception fusion center FC is received is N₁×N₂It is individual；

(3) frequency spectrum perception fusion center FC is according to primary election time user CR'_tConfidence level P'_t, business demand R'_t, primary election time is used Family CR'_tDistribution needs the number of frequency bands C detected_t：

(3-1) is according to each primary election time user CR'_tConfidence level P'_t, respectively to each primary election time user CR'_tConfidence level P'_t It is normalized, obtains each primary election time user CR'_tNormalization confidence value

$\begin{array}{cc}\stackrel{\‾}{{P^{\′}}_t}=\frac{{P^{\′}}_t}{\underset{t=1}{\overset{{N^{\′}}_2}{\Σ}}{P^{\′}}_t},& 1\≤t\≤{N^{\′}}_2;\end{array}$

(3-2) each primary election time user CR' according to obtained by step (3-1)_tCorresponding normalization confidence valueCalculate frequency spectrum sense Know fusion center FC distribution primary election time users CR'_tNeed the number of frequency bands C of detection_t：

$\begin{array}{cc}{C}_t=\stackrel{\‾}{{P^{\′}}_t}\·{N^{\′}}_2,& 1\≤t\≤{N^{\′}}_2;\end{array}$

(4) frequency spectrum perception fusion center FC is according to the primary election time user CR' for participating in cooperative detection_tSignal to noise ratio snr '_t, calculate institute There is the signal to noise ratio root-mean-square value of primary election time userAnd make signal to noise ratio snr '_t=γ_t, wherein, signal to noise ratio root-mean-square valueCalculating It is as follows：

$\begin{array}{cc}\stackrel{\‾}{\mathrm{\γ}}=\sqrt{\frac{1}{{N^{\′}}_2}\underset{t=1}{\overset{{N^{\′}}_2}{\Σ}}{\left({{\mathrm{SNR}}^{\′}}_t\right)}^2},& {N^{\′}}_2\≤N_2;\end{array}$

(5) frequency spectrum perception fusion center FC calculates the signal to noise ratio root-mean-square value of all primary election time user successively respectivelyWith each primary election Secondary user CR'_tSignal to noise ratio snr '_tBetween quotient η_t, wherein,

$\begin{array}{ccc}{\mathrm{\η}}_t=|\stackrel{\‾}{\mathrm{\γ}}/{\mathrm{\γ}}_t|,& t=1,2,...,{N^{\′}}_2,& {N^{\′}}_2\≤N_2;\end{array}$

(6) frequency spectrum perception fusion center FC is calculated, is obtained signal to noise ratio predetermined threshold value λ and signal to noise ratio optimal threshold λ_optimal, and respectively According to each signal to noise ratio quotient η_tWith the magnitude relationship between signal to noise ratio predetermined threshold value λ, the final election time user CR for participating in cooperation is selected ”_k, final election time user CR "_kSignal to noise ratio be SNR "_k, wherein,

(6-1) frequency spectrum perception fusion center FC is according to the N' of reception₂Individual primary election is from the corresponding signal to noise ratio set { SNR' of user_t, obtain Primary election is taken from user's signal to noise ratio set { SNR'_tIn signal to noise ratio maximum, remember the signal to noise ratio maximum be SNR'max；

(6-2) using the signal to noise ratio maximum SNR'max of acquisition as reference, and by N'₂Individual primary election is from user CR'_tSignal to noise ratio SNR'_tMake business's processing with signal to noise ratio maximum SNR'max respectively, calculating obtain each primary election from user's signal to noise ratio snr '_tCorresponding Initial threshold λ_t, wherein,

λ_t=| SNR'_t/SNR'_max|, t=1,2 ..., N'₂, N'₂≤N₂；

(6-3) is according to each primary election from user CR'_tNormalization confidence valueWith signal to noise ratio quotient η_t, each primary election is calculated from user CR'_tJoint screening parameter value ξ_t, and according to joint screening parameter value ξ_t, the final election for participating in cooperation is chosen from user CR "_k, its In, final election is from user CR "_kQuantity be M,T=1,2 ..., N '₂, k=1,2 ..., M, M≤N'₂：

If joint screening parameter value ξ_tPositioned at default value interval range [ξ_a,ξ_b] in, i.e. ξ_a≤ξ_t≤ξ_bWhen, then choose the joint Screening parameter value ξ_tCorresponding primary election from user be final election from user, and participate in cooperative detection；Otherwise, the primary election is refused from user Choose；

The signal to noise ratio predetermined threshold value λ of (6-4) in step (6-3), obtains M final election from user CR "_kRespectively in OR criterions and Cooperative detection performance curve under AND criterions, wherein,

OR criterions：

AND criterions：

$\begin{array}{cc}{P}_{d,k}=\frac{\underset{k}{\overset{{N^{\′}}_2}{\Σ}}\underset{s=1}{\overset{C_t}{\Σ}}{P}_{d,ks}}{{N^{\′}}_2\·C_t},& P_{fa,k}=\frac{\underset{k}{\overset{{N^{\′}}_2}{\Σ}}\underset{s=1}{\overset{C_t}{\Σ}}{P}_{fa,ks}}{{N^{\′}}_2\·C_t};\end{array}$

Wherein, P_d,kIt is k-th of final election from user CR "_kAverage detected probability, P_fa,kIt is k-th of final election from user CR "_kIt is flat Equal false-alarm probability；P_d,ksIt is final election from user CR "_kTo the detection probability of its allocated s-th of frequency range, P_fa,ksIt is final election from user CR”_kTo the false-alarm probability of its allocated s-th of frequency range；Q_dFor the global detection probability after cooperative detection, Q_faAfter cooperative detection Global false-alarm probability；ω_kRepresent signal to noise ratio CR "_kWeight coefficient, SNR "_kIt is k-th of final election from user CR "_kSignal to noise ratio, SNR”_maxRepresent signal to noise ratio maximum of the M final election from user, SNR "_minRepresent signal to noise ratio minimum value of the M final election from user；

(6-5) is respectively obtained under OR criterions and AND criterions according to the cooperative detection performance curve under OR criterions and AND criterions Maximum detection probability Q_{(OR, d)-max}、Q_{(AND, d)-max}, obtain Q_{(OR, d)-max}And Q_{(AND, d)-max}Maximum Q_d-max, and it is optimal with this Detect performance number Q_d-maxCorresponding signal to noise ratio predetermined threshold value is signal to noise ratio optimal threshold, and note signal to noise ratio optimal threshold is λ_optimal；Wherein, Q_d-max=max (Q_{(OR, d)-max}, Q_{(AND, d)-max})；

(7) according to the signal to noise ratio optimal threshold λ of acquisition_optimal, obtain signal to noise ratio optimal threshold λ_optimalCorresponding final election from Family CR ", obtains the final election from user CR " Dynamic gene α and other M-1 final election from user CR "_kDynamic gene α_k, and Respectively according to Dynamic gene α, α_kCorrespondence adjusts final election from user CR ", CR "_kAverage false-alarm probability, final election from user CR " adjust Average false-alarm probability afterwards is designated as P_fa, final election is from user CR "_kAverage false-alarm probability after adjustment is designated as P_fa,k；Wherein,

P_fa,k=α_k·P_fa, k=1,2 ..., M-1；

$\begin{array}{cc}{\mathrm{\α}}_k=1+\frac{\stackrel{\‾}{{\mathrm{SNR}}^{\′}}-{{\mathrm{SNR}}^{\′\′}}_k}{\stackrel{\‾}{{\mathrm{SNR}}^{\′\′}}},& k=1,2,...,M-1;\end{array}$

$\begin{array}{cc}\stackrel{\‾}{{\mathrm{SNR}}^{\′\′}}=\sqrt{\underset{k=1}{\overset{M}{\Σ}}\frac{{\left({{\mathrm{SNR}}^{\′\′}}_k\right)}^2}{M}},& M\≤{N^{\′}}_2;\end{array}$

Wherein, α_kIt is final election from user CR "_kDynamic gene, for according to final election from user CR "_kThe signal to noise ratio snr of itself "_kIt is real Now to the adjustment of its average false-alarm probability size；SNR”_kIt is k-th of final election from user CR "_kSignal to noise ratio；

(8) according to the M final election obtained in step (7) from the Dynamic gene α of user_kAnd the average false-alarm after correspondence adjustment is general Rate P_fa,k, final election is calculated from user CR "_kJudgement threshold values λ " after adjustment_kWith average detection probability P_d,k, wherein,

$\begin{array}{c}\mathrm{\λ}={\mathrm{\σ}}_w^2\[\sqrt{2n}{Q}^{-1}\left(P_{fa,k}\right)+n\]\\ ={\mathrm{\σ}}_w^2\[\sqrt{2n}{Q}^{-1}(\mathrm{\δ}\·P_{fa})+n\]\\ ={\mathrm{\σ}}_w^2\[\sqrt{2n}{Q}^{-1}\left(\right(1+\frac{\stackrel{\‾}{{\mathrm{SNR}}^{\′\′}}-{{\mathrm{SNR}}^{\′\′}}_k}{\stackrel{\‾}{{\mathrm{SNR}}^{\′\′}}})\·P_{fa})+n\]\end{array};$

$P_{d,k}=Q\[Q^{-1}\left(P_{fa,k}\right)-\sqrt{n\·{{\mathrm{SNR}}^{\′\′}}_k}\];$

$n=2{\[Q^{-1}\left(P_{fa,k}\right)-Q^{-1}\left(P_{fa}\right)\sqrt{1+2{{\mathrm{SNR}}^{\′\′}}_k}\]}^2\·{\left({{\mathrm{SNR}}^{\′\′}}_k\right)}^{-2};$

Wherein,N is sampling number；

(9) according to signal to noise ratio snr of the M final election from user in step (8) "_kAnd the average detected probability after obtained adjustment P_d,k, return to step (6) selects in M final election from user again, and obtain participating in cooperation T selects from user CR " ' eventually_t, and Using the global detection probability after the OR criterions cooperation of weighting as frequency spectrum perception fusion center FC final detection result, wherein 1≤t ≤T≤M≤N'₂；Wherein：

The OR criterions of the weighting are as follows：

$\begin{array}{cc}{Q^{\′}}_d=1-\underset{t=1}{\overset{M^{\′}}{\Π}}{{\mathrm{\ω}}^{\′}}_t(1-{P^{\′}}_{d,t}),& {Q^{\′}}_{fa}=1-\underset{t=1}{\overset{M^{\′}}{\Π}}{{\mathrm{\ω}}^{\′}}_t(1-{P^{\′}}_{f,t});\end{array}$

$\begin{array}{ccccc}{{\mathrm{\ω}}^{\′}}_t=\frac{{P^{\′}}_{d,t}}{\underset{t=1}{\overset{M^{\′}}{\Σ}}{P^{\′}}_{d,t}},& {P^{\′}}_{d,t}=\frac{\underset{t}{\overset{M^{\′}}{\Σ}}\underset{s=1}{\overset{C_t}{\Σ}}{P^{\′}}_{d,ts}}{M^{\′}\·C_t},& {P^{\′}}_{fa,t}=\frac{\underset{t}{\overset{M^{\′}}{\Σ}}\underset{s=1}{\overset{C_t}{\Σ}}{P^{\′}}_{fa,ts}}{M^{\′}\·C_t},& t=1,2,\mathrm{...},M^{\′},& M^{\′}\≤M;\end{array}$

Wherein, P'_d,tsTo select eventually from user CR " '_tTo the detection probability of its allocated s-th of frequency range, P'_fa,tsTo select eventually from user CR”'_tTo the false-alarm probability of its allocated s-th of frequency range；P'_d,tSelected for the end that t-th reselects from user CR " '_tBe averaged Detection probability, P'_fa,tSelected for the end that t-th reselects from user CR " '_tAverage false-alarm probability；Q'_dAfter cooperative detection Global detection probability, Q'_faFor the global false-alarm probability after cooperative detection；M' is the whole number selected from user reselected； ω'_tIt is the whole choosing that reselects from user CR " '_tWeight coefficient.