CN103200670B - The cognitive radio primary user localization method of convex set projection is checked based on backtracking - Google Patents

Abstract

The present invention proposes a method for locating the main user of cognitive radio based on retrospectively checking the convex set projection. Taking the coordinates of L sensing users as the center of the circle, the convex set projection algorithm is used to perform Mc-step orthogonal projection iteration on the main user; perform mc-step backtracking Compare and check, calculate the distance between adjacent iteration points, and compare it with the threshold value λ; if the distance between adjacent iteration points is partially zero or greater than λ, then carry out the convex set circle domain boundary Mb step projection iteration, compare and check with mb step backtracking, calculate the distance value between two adjacent iteration points, and compare it with the threshold value λ again, if both are less than λ, then determine the iteration result b _Mb of Mb step as the main The positioning result of the user's location information. The retrospectively checking convex set projection positioning algorithm in the present invention makes up for the deficiency of the existing convex set projection positioning algorithm, the positioning algorithm is good, and is less affected by the ranging error, and is suitable for the perception of the main user in the cognitive radio network. Obtaining link of location information.

Description

Cognitive Radio Primary User Location Method Based on Backchecking Convex Set Projection

technical field

The invention relates to a method for locating a primary user in a cognitive radio network, in particular to a convex set projection locating method based on backtracking inspection.

Background technique

With the combination of wireless mobile communication and computer network application development becoming more and more mature, the mobile Internet has become one of the fastest growing businesses with the greatest market potential and the highest commercial value in the world today. Rich applications mainly rely on information carrying methods such as text, images, and videos, and with people's continuous pursuit of application quality, information transmission is required to be more efficient and convenient. The innovation and development of these applications require wider spectrum and higher download rate. The concept of cognitive radio caters to the needs of spectrum reuse. It can avoid conflicts with the primary user through the perception of the wireless environment, and use optimal decision-making to effectively and dynamically utilize spectrum holes. If the location information of the primary user can be obtained, the performance of spectrum sensing will be greatly improved, and it will also play a great role in the subsequent management and allocation of spectrum resources.

To locate the main user in the cognitive network, the main functions of obtaining the location information of the main user are as follows:

1. Provide support for spectrum resource management. When the location information of the primary user is known, the spectrum utilization can be better improved according to the location information, and the sensing user can be better guided not to interfere with the spectrum usage of the primary user.

2. Reduce the power consumption of users in the cognitive network. When the location information of the primary user is known, the sensing user in the cognitive network can determine the direction of spectrum sensing according to the location information of the primary user, and can accurately judge the spectrum usage of the primary user in the minimum power operating state .

3. Avoid interference to the primary user. When the position information of the primary user is known, the multi-antenna technology can be combined to perform spectrum sensing for the direction and position of the primary user, avoiding the possibility of mutual interference between spectrums.

4. It is conducive to the location optimization of the perceived user. When the location information of the primary user is known, according to the location information of the primary user, the location of the sensing users can be reasonably distributed, the utilization rate of spectrum and space can be improved, and the interference to the primary user can be better avoided.

At present, the commonly used convex set projection methods include Circular POCS, Hyperbolic POCS, BoundaryPOCS and Hybrid POCS, among which Hybrid POCS is the combination of the first two POCS methods. According to the research results, the positioning accuracy of the Hybrid POCS method is better than that of the previous methods. , however, when the main user is far away from the sensing user, because the hyperbolic projection positioning in the Hybrid POCS algorithm is subject to large noise fluctuations for the main user when the hyperbolic projection positioning is outside the sensing user polygon, the error increases as the distance increases .

Contents of the invention

The present invention aims to solve the above-mentioned technical defects, and proposes a backcheck convex set projection algorithm (BackCheck POCS) applied to the positioning of the main user in the cognitive network.

The method includes the following steps:

Step 1. Taking the coordinates of L sensing users as the center of the circle, use the convex set projection algorithm to perform Mc-step orthogonal projection iterations on the main users, and obtain Mc iteration points x _k , where k=1,2,3,...,Mc;

Step 2. For the Mc iteration points obtained in step 1, perform mc step backtracking comparison check, and calculate the distance ||x _m+1 -x _m || between adjacent iteration points, where m=Mc-1 ,...,Mc-mc

Step 3. If in the backtracking comparison check in step 2, the distance between adjacent iteration points is less than λ and not zero, then the last L iteration mean value in step 1 is used as the positioning result of the main user position information; if step In the backtracking comparison check in the second step, if the distance between adjacent iteration points is partially zero or greater than λ, proceed to step four;

Step 4. Take the Mc step iteration result x _Mc as the initial point b ₀ , perform orthogonal projection iteration on the boundary of the convex set circular domain, and the number of iteration check steps is Mb, and obtain Mb iteration points b _h , where h=1,2, 3,... Mb;

Step 5. For the Mb iteration points obtained in step 4, perform mb-step backtracking comparison checks, and calculate the distance value ||b _n+1 -b _n || between adjacent two iteration points, where n=Mb- 1,..., Mb-mb, and compare with the threshold value λ;

Step 6. If in the backtracking comparison check in step 5, the distance values between adjacent iteration points are all less than λ, then the last L iteration mean values in step 4 are used as the positioning result of the main user position information; if the backtracking in step 5 In the comparison check, if the distance value between adjacent iteration points is greater than λ, skip to step 4, take Mb step iteration result b _Mb as the initial point b ₀ , and change the projection iteration order until the distance between adjacent iteration points The values are all less than λ.

Preferably, said step one includes:

1.1) Initialization step: set the initial point x ₀ , where x ₀ is a point at any position;

1.2) Use the following formula for projection iteration:

in, Represents an orthogonal convex set projection point, Represents the vector from P _i to P _i+1 ; P _i is the location coordinate of the i-th perceived user, i∈[1,L]. D _i is a convex circular domain with the i-th sensing user as the center and the distance measured between the i-th sensing user and the main user as the radius.

Preferably, the value of λ depends on the average value of the ranging results of the sensing user to the primary user, and λ is a small value relative to the average distance.

Preferably, mb and mc take the same value, which is an integer multiple of L.

Preferably, said step four includes:

4.1) Initialization steps: 1) Set the initial point b ₀ , b ₀ = x _Mc

4.2) Use the following formula for projection iteration:

b _h+1 = _{Ph mod L} (b _h ), h = 0, 1, 2, 3... Mb-1

in, $P_i\left(\mathrm{the\; y}\right)=\arg \underset{x\∈C_i}{\min }\left|\right|\mathrm{the\; y}-x\left|\right|=P_i+d_i*\frac{\mathrm{the\; y}-P_i}{\left|\right|\mathrm{the\; y}-P_i\left|\right|}$

Among them, P _i is the location coordinate of the i-th perceived user, i∈[1,L]. d _i is the measured distance between the i-th sensing user and the main user; C _i ={y∈R ² :||yP _i ||=d _i } is the radius determined by the i-th sensing user is the circle boundary of d _i .

This algorithm is based on the improvement of the convex set projection positioning algorithm, which makes up for the shortcomings of the existing convex set projection positioning algorithm, and is relatively less affected by the ranging error, and is suitable for the acquisition of the location information of the primary user by the sensing user in the cognitive radio network. Link, can more accurately realize the positioning of the main user.

Description of drawings

Fig. 1 is the flowchart based on BackCheck POCS positioning method among the present invention.

Fig. 2 is the iterative schematic diagram of the present invention based on the BackCheck POCS positioning method.

Figure 3 is a schematic diagram of the positioning error comparison between BackCheck POCS and Hybrid POCS.

Figure 4 is a schematic diagram of the comparison of the influence of different ranging errors on the positioning error of BackCheck POCS and Hybrid POCS.

The embodiments described below by referring to the figures are exemplary only for explaining the present invention and should not be construed as limiting the present invention.

Assuming that there are L sensing users participating in the positioning of the main user, the location coordinates of the L sensing users are known, expressed as:

${<span\; class="notranslate">\; \{</span>{\mathrm{<span\; class="notranslate">\; P</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; Y</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Element;</span>{\mathrm{<span\; class="notranslate">\; R</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \}</span>}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; L</span>}}$

The measured distance between the L sensing users and the primary user is expressed as:

${<span\; class="notranslate">\; \{</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; \}</span>}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; L</span>}}$

Taking each perceived user as the center and taking the distance measurement value d _i as the radius of the convex set circle domain is expressed as:

${<span\; class="notranslate">\; \{</span>{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Subset;</span>{\mathrm{<span\; class="notranslate">\; R</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \}</span>}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; L</span>}}$

The circular boundary with radius d _i determined by the i-th perception user is expressed as follows:

C _i ={y∈R ² :||yP _i ||=d _i }

The POCS algorithm used in the positioning method of the present invention can be Circular POCS, Hyperbolic POCS and Boundary POCS. Taking the Circular POCS algorithm as an example, the specific steps based on the BackCheck POCS algorithm are given:

Step 1. Taking the coordinates of the L perceived users as the center of the circle, the Circular POCS is used to perform Mc-step orthogonal projection iterations on the primary user to obtain Mc iteration points x ₁ , x ₂ , ... x _Mc-1 , x _Mc .

Here, the number of iterative inspection steps of Circular POCS is set as Mc. Since the convergence speed of Circular POCS is very fast, the value of Mc can be appropriately taken as a small value.

According to the projection iteration rules of Circular POCS, the steps of Mc step orthogonal projection iteration for the main user are as follows:

1) Initialization: set the initial point x ₀ , where x ₀ is any point in the plane, as shown in the square in Figure 2;

2) Orthogonal projection iteration of simplified circular POCS:

in Represents an orthogonal convex set projection point, Represents the vector from P _i to P _i+1 , and i can indicate the order of iterations, which is determined according to the order of perceived users; when subsequent iterations enter an infinite loop or fail to converge, the order of iterations can be changed and iterations continued.

Step 2. For the Mc iteration points obtained in step 1, perform mc step backtracking inspection. Calculate the distance ||x _m+1 -x _m || between adjacent iteration points, where m=Mc-1,...,Mc-mc, and compare it with the threshold value λ. Among them, the value of λ depends on the average value of the ranging results of the sensing user to the primary user, and λ is a small value relative to the average distance, for example: the ratio of λ to the average distance is less than or equal to 0.02, considering the algorithm The computational complexity can further limit the ratio within 0.005~0.02.

The number of backtracking inspection steps is mc, and the value of mc is an integer multiple of the number L of sensing users participating in positioning.

Step 3. If the distance between adjacent iteration points is less than λ and not zero in the backtracking comparison check in step 2, it can be judged that the main user is located within the polygon formed by the perceived user, and the last L in step 1 iteration mean, that is, the mean value of the last iteration of each convex set to which the perceived user belongs As the positioning result of the position information of the primary user; if the distance between adjacent iteration points is partly zero or greater than λ in the retrospective comparison check in step 2, proceed to step 4.

Step 4. Taking the iteration result x _Mc of step Mc as the initial point b ₀ , perform orthogonal projection iteration on the boundary of the convex set circular domain, and the number of iteration checking steps is Mb, and Mb iteration points are obtained.

Set the number of inspection steps of the boundary orthogonal projection iteration as Mb. Since the judgment of the position of the iterated point is cancelled, the convergence speed of the boundary orthogonal projection iteration is uncertain, and it may quickly converge to the vicinity of the main user, or may fall into The cycle is slow, so Mb takes a larger value to make the iteration of the boundary orthogonal projection sufficient.

Among them, according to the rule of boundary orthogonal projection iteration, the iteration steps of Mb steps for the main user are:

1) Set the initial point b ₀ , b ₀ = x _Mc

2) b _h+1 = P _hmodL (b _h ), h=0, 1, 2, 3...Mb-1

in, $P_i\left(\mathrm{the\; y}\right)=\arg \underset{x\&Element;C_i}{\min }\left|\right|\mathrm{the\; y}-x\left|\right|=P_i+d_i*\frac{\mathrm{the\; y}-P_i}{\left|\right|\mathrm{the\; y}-P_i\left|\right|}$

Step 5. Perform a backtracking mb step check on the Mb iteration points obtained in step 4. Calculate the distance value between two adjacent iteration points, and compare it with the threshold value λ.

The number of backtracking inspection steps is mb, and the value of mb is the same as mc, which is an integer multiple of L.

Step 6. If in the retrospective comparison check in step 5, the distance values between adjacent iteration points are all less than λ, then the mean value of the last L iterations in step 4, that is, the mean value of the last iteration of each convex set to which the perceived user belongs As the positioning result of the position information of the main user; if the distance value between adjacent iteration points is greater than λ in the retrospective comparison check in step 5, it means that the boundary orthogonal projection still has no converges to near the main user position, but gets stuck in a slow loop iteration. At this time, jump to step 4 and continue to perform the iteration of orthogonal projection on the boundary of the circular domain of the convex set, where the iteration result b _Mb of step Mb in the previous Mb iteration of orthogonal projection is the initial point b ₀ , and transform the original projection Iteration order until the distance values between adjacent iteration points are less than λ.

The present invention will be described in further detail below in conjunction with the accompanying drawings and specific examples.

Step 1. Taking the coordinates of the L sensing users as the center of the circle, the Circular POCS is used to perform Mc-step orthogonal projection iterations on the main user, and Mc iteration points x _k are obtained.

Set the number of sensing users as L=3, and the position coordinates of sensing users as [(700m1500m), (500m1000m), (1000m, 1000m)]. Wherein, white noise is input as the measured value d _i of the distance between the sensing user and the main user, and the variance is 8m. Set the upper line Mc=10 of the projection iteration steps of the circular convex set, take the coordinate position x ₀ = (1600m, 2100m) as the starting position (as shown in the square in Figure 2) to perform projection iterations, and obtain 10 iteration points x _k , k =1,2,3,...10.

Step 2. For the 10 iteration points obtained in step 1, perform mc-step backtracking inspection, and the number of backtracking inspection steps is mc=2L=6. Calculate the distance between adjacent iteration points and compare with the threshold value λ. Assuming that the average value of the distance measurement values obtained by all sensing users participating in positioning and the main user is R, the check threshold value λ is set to be a small amount relative to R, here we set λ and the average distance The ratio of λ/R=0.01.

||x _m+1 -x _m ||≤λ, where, m=Mc-1,...,Mc-mc

Step 3. Since the distance between adjacent iteration points is partially zero or greater than λ in the backtracking comparison check in step 2, proceed to step 4.

It can be seen from Figure 2 that after two iterations of Circular POCS starting from the initial point x ₀ , the iteration point stagnates at the intersection area of three convex circular domains. Part of the change difference is zero, so proceed to step 4 to iterate the orthogonal projection onto the boundary of the convex set circular domain.

Step 4. Taking the iteration result x _Mc of step Mc as the initial point, perform orthogonal projection iterations on the boundary of the convex set circular domain, perform Mb steps of orthogonal projection iterations on the main user, and obtain Mb iteration points b _h , h=1, 2,3,...Mb.

First, set the upper limit of the iterative steps of the boundary orthogonal projection Mb=30. Since the judgment of the location of the iterated point is canceled, the convergence speed of the boundary orthogonal projection iteration is uncertain, and it may converge to the vicinity of the main user very quickly. may fall into a slow cycle, so Mb takes a larger value to make the boundary orthogonal projection iteration sufficient; then, according to the boundary orthogonal projection iteration rule, Mb step positioning projection is performed on the main user to obtain 10 iteration points b _h , h= 1, 2, 3, ... 30.

Step 5. Perform a backtracking mb step check on the Mb iteration points obtained in step 4. The number of steps for backtracking checks mb=2L=6. Calculate the distance value between two adjacent iteration points, and compare it with the threshold value λ.

Step 6. After iterating the Mb step, backcheck the change value of the iteration point at the mb=6 step, and find that the distance value between adjacent iteration points is greater than λ, and the projection iteration falls into a slow loop projection iteration. Therefore, it is necessary to skip Go to step 4, and take the iteration result of the Mb-th step in the first Mb-step orthogonal projection iteration as the initial point b ₀ , execute the Mb-step orthogonal projection iteration on the boundary of the circular domain of the convex set again, and change the first projection iteration order, change the order of the original projection iteration from P1-P2-P3 to P2-P3-P1, after Mb steps of iteration again, check mb=6 steps back, and find that the change value of the iteration point is less than the threshold λ, indicating that the iteration converges mainly The location of the user is nearby, so the mean value of the last L iterations in the second Mb-step orthogonal projection iteration (as shown by the asterisk in Figure 2) is used to determine the location of the main user.

Figure 3 is a comparison of the positioning simulation results between the Hybrid POCS positioning algorithm and the BackCheck POCS positioning algorithm. The abscissa in the figure is the number of simulation repetitions, and the ordinate is the ratio of the difference between the estimated position and the real target position to the average value of the real distance between the perceived user and the main user. It can be seen from Figure 3 that in general, the positioning accuracy of the two algorithms is relatively close, but in some cases, the positioning accuracy of the BackCheckPOCS positioning algorithm has an advantage. This is because, when the main user is far away from the sensing user, the asymptotic property of the hyperbola tends to cause the intersection point of the hyperbola to be significantly affected by the fluctuation of the ranging noise. Therefore, in this case, the positioning effect of the BackCheck POCS algorithm can be seen It is superior to the HybridPOCS algorithm.

Figure 4 describes the comparison of the positioning accuracy between the Hybrid POCS positioning algorithm and the BackCheck POCS positioning algorithm under the influence of different ranging errors. It can be seen from Figure 4 that the BackCheck POCS positioning algorithm has certain advantages over the Hybrid POCS positioning algorithm, which is mainly due to the fact that the hyperbolic projection positioning in the Hybrid POCS algorithm is affected by noise for the main user when the main user perceives the user polygon. The fluctuation is large, so as the ranging error increases.

Although the embodiments of the present invention have been shown and described, those skilled in the art can understand that various changes, modifications and substitutions can be made to these embodiments without departing from the principle and spirit of the present invention. and modifications, the scope of the invention is defined by the appended claims and their equivalents.

Claims (4)

1. A method for primary user location in a cognitive radio network, characterized in that the method comprises the following steps: Step 1. Taking the coordinates of the L perceived users as the center of the circle, use the convex set projection algorithm to perform Mc step orthogonal projection iterations on the main user, and obtain Mc iteration points x _k, where k=1, 2, 3, ..., Mc; Step 2. For the Mc iteration points obtained in step 1, carry out mc step backtracking comparison check, calculate the distance ||x _m+1 -x _m || between adjacent iteration points, wherein, m=Mc-1 ,...,Mc-mc, and compare with the threshold value λ; Step 3. If in the backtracking comparison check in step 2, the distance between adjacent iteration points is less than λ and not zero, then the last L iteration mean value in step 1 is used as the positioning result of the main user position information; if step In the backtracking comparison check in the second step, if the distance between adjacent iteration points is partially zero or greater than λ, proceed to step four; Step 4. Take the Mc step iteration result x _Mc as the initial point b ₀ , perform orthogonal projection iteration on the boundary of the convex set circular domain, and the number of iteration check steps is Mb, and obtain Mb iteration points b _h , where h=1,2, 3,...,Mb; Step 5. To the Mb iteration points obtained in step 4, carry out mb step backtracking comparison check, calculate the distance value ||b _n+1 -b _n || between adjacent two iteration points, wherein, n=Mb- 1,...,Mb-mb, and compare with the threshold value λ; Step 6. If in the backtracking comparison check in step 5, the distance values between adjacent iteration points are all less than λ, then the last L iteration mean values in step 4 are used as the positioning result of the main user position information; if the backtracking in step 5 In the comparison check, if the distance value between adjacent iteration points is greater than λ, skip to step 4, take Mb step iteration result b _Mb as the initial point b ₀ , and change the projection iteration order until the distance between adjacent iteration points The values are all less than λ; Described step one comprises: 1.1) Initialization step: set the initial point x ₀ , where x ₀ is a point at any position; 1.2) Use the following formula for projection iteration: , k=0,1,2...Mc-1 in, Represents an orthogonal convex set projection point, Indicates the vector from P _i to P _i+1 ; P _i is the location coordinate of the _i -th sensing user, i∈[1,L]; The obtained distance measurement value from the main user is a convex circle domain with a radius. 2. The method for locating the primary user in the cognitive radio network according to claim 1, wherein the value of λ depends on the average value of the distance measurement results of the cognitive user to the primary user, and λ is relative to the average The distance is a small value. 3. The method for locating a primary user in a cognitive radio network as claimed in claim 1, wherein mb and mc take the same value, which is an integer multiple of L. 4. The method for locating a primary user in a cognitive radio network according to claim 1, wherein said step 4 comprises: 4.1) Initialization steps: 1) Set the initial point b ₀ , b ₀ =x _Mc 4.2) Use the following formula for projection iteration: b _h+1 ＝P _{h mod L} (b _h ), h=0, 1, 2, 3...Mb-1 Among them, let h mod L be i, b _h be y, then $P_i\left(\mathrm{the\; y}\right)=\arg \underset{x\&Element;C_1}{\min }\left|\right|\mathrm{the\; y}-x\left|\right|=P_i+d_i*\frac{\mathrm{the\; y}-P_i}{\left|\right|\mathrm{the\; y}-P_i\left|\right|}$ Among them, P _i is the location coordinates of the i-th sensing user, i∈[1,L]; d _i is the measured distance between the i-th sensing user and the main user; C _i ={y∈R ² :||yP _i ||=d _i } is the circle boundary with radius d _i determined by the i-th sensing user.