A multi-channel collaborative spectrum sensing method and system

Abstract

A multi-channel collaborative spectrum sensing method and system, specifically relates to a method and system for acquiring collaborative spectrum sensing between multiple nodes in a computer cognitive wireless network under a multi-channel situation. In order to solve the problems of low accuracy and poor real-time performance of multi-channel collaborative spectrum sensing, it is defined that primary users and secondary users in a cognitive wireless network obey homogeneous Poisson point process distribution respectively; each secondary user performs energy detection on each primary user channel in turn to obtain a detection probability matrix; clustering restriction conditions are set, and primary users and secondary users are clustered using a bipartite graph; a secondary user with the largest average signal-to-noise ratio of primary user signals received in each cluster is selected as a cluster head, the cluster head collects spectrum sensing information of other secondary users in the cluster, and fuses it using an OR fusion rule, and judges whether the primary user channel corresponding to each cluster has a primary user signal based on the fused spectrum sensing information, and fuses it with other cluster heads to obtain the occupancy status of all primary user channels.

Description

Technical Field

[0001] The present invention relates to a collaborative spectrum sensing method and system, and in particular to a method and system for acquiring collaborative spectrum sensing between multiple nodes in a computer cognitive wireless network under a multi-channel situation, and belongs to the field of wireless communications. Background Art

[0002] Collaborative spectrum sensing technology is a technology that involves multiple nodes working together to sense the spectrum and integrate the sensing results of multiple nodes. It can effectively resist external interference and reduce the requirements for the accuracy of spectrum sensing of a single node. Collaborative spectrum sensing technology mainly involves three processes: local sensing, information transmission, and data fusion. Among them, local sensing is the basis of collaborative spectrum sensing technology. Multi-channel collaborative spectrum sensing is to expand the single-channel situation to the multi-channel situation. By dividing a wider frequency band into multiple narrower frequency bands, these narrower frequency bands are sensed separately, so as to achieve the purpose of detecting the spectrum occupancy of a wider frequency band.

[0003] Cognitive Radio (CR) is an intelligent wireless communication technology that improves spectrum utilization. It allows secondary users (also known as cognitive users) to detect the authorized frequency band (or authorized channel) used by the primary user (or authorized user) in real time. When the secondary user detects that the authorized channel is idle, the secondary user adjusts its own transmission power and signal modulation parameters, so that the authorized channel assigned to the primary user can be used for communication while ensuring the communication quality of the primary user. When the primary user needs to occupy the authorized channel of a primary user, the secondary user must exit from the authorized channel and search and detect other idle channels to complete its own communication. The communication network where the primary user and the secondary user are located is a cognitive wireless network. With the increase in the number of nodes in cognitive wireless networks, the topology of the network is becoming more and more complex. In the problem of multi-channel spectrum sensing, if the single-channel spectrum sensing method is adopted to use all nodes to perform spectrum sensing on a single sub-channel one by one, it will produce obvious sensing delay, which is difficult to meet the real-time requirements of spectrum sensing and difficult to reflect the dynamic changes in spectrum usage. Therefore, most traditional methods mainly allocate perception tasks to a single node. However, this collaborative perception strategy will bring a large system overhead in the process of transmitting local perception results, and the bandwidth requirements of the communication system are also relatively high, resulting in reduced efficiency of spectrum perception, low accuracy and poor real-time performance. Summary of the invention

[0004] In order to solve the problem that the current spectrum sensing method mainly allocates sensing tasks to a single node, which will bring a large system overhead in the process of local sensing result transmission and has a relatively high bandwidth requirement for the communication system, resulting in low accuracy and poor real-time performance of multi-channel collaborative spectrum sensing, the present invention further proposes a multi-channel collaborative spectrum sensing method and system.

[0005] It includes the following steps:

[0006] S1. Acquire a cognitive wireless network, obtain primary user information and secondary user information in the cognitive wireless network, wherein the primary user and the secondary user obey homogeneous Poisson point process distributions in the cognitive wireless network, and the two homogeneous Poisson point processes are independent of each other;

[0007] S2. In the local sensing stage of collaborative spectrum sensing, each secondary user performs energy detection on each primary user channel in turn to obtain a detection probability matrix;

[0008] S3. In the information transmission stage of collaborative spectrum sensing, set clustering restriction conditions, and cluster the primary users and secondary users using a bipartite graph according to the restriction conditions and the detection probability matrix to obtain multiple clusters;

[0009] S4. In the data fusion stage of collaborative spectrum sensing, the secondary user with the largest average signal-to-noise ratio of the primary user signal received in each cluster is selected as the cluster head. The cluster head collects the spectrum sensing information of other secondary users in the cluster, and fuses the spectrum sensing information of other secondary users using the OR fusion rule. It is determined whether there is a primary user signal in the primary user channel corresponding to the cluster based on the fused spectrum sensing information, and the existence of the primary user signal is interactively fused with other cluster heads to obtain the occupancy status of all primary user channels.

[0010] S5. Obtain the cognitive wireless network to be identified, and repeat S1-S4 to obtain the occupancy status of all primary user channels.

[0011] Furthermore, the primary user information in S1 includes the number of primary users and the position of each primary user in the cognitive wireless network, and the secondary user information includes the number of secondary users and the position of each secondary user in the cognitive wireless network.

[0012] Furthermore, the homogeneous Poisson point process in S1 is:

[0013] When a point process in a certain space in a cognitive wireless network satisfies the following two conditions at the same time, it is called a homogeneous Poisson point process;

[0014] Condition 1: For any bounded area D in a cognitive wireless network, the number of nodes N(D) in the area obeys a Poisson distribution;

[0015] Condition 2: There are disjoint bounded regions D in cognitive wireless networks 1 ,D 2 , ..., D n , the total number of nodes in the regions are independent of each other.

[0016] Furthermore, in the local sensing process of the collaborative spectrum sensing in S2, each secondary user performs energy detection on each primary user channel in turn to obtain a detection probability matrix. The specific process is:

[0017] The secondary user processes the received primary user signal through a bandpass filter, a square-law element, an analog-to-digital converter, an integrator or an accumulator in sequence to obtain an energy detection statistic, and compares the energy detection statistic with a custom threshold value. If the energy detection statistic is greater than the threshold value, it indicates that the primary user signal exists, and the detection probability of the primary user is calculated to obtain a detection probability matrix. The detection probability matrix includes the detection probability of each secondary user for each primary user channel; if the energy detection statistic is less than the threshold value, it indicates that the primary user signal does not exist.

[0018] Furthermore, the energy detection statistic is:

[0019]

[0020] Among them, Y i,j represents the energy detection statistic of the i-th secondary user to the j-th primary user signal; f s is the sampling frequency; τ is the sampling time; y i,j (n) represents the binary perception model of the i-th secondary user to the j-th primary user channel.

[0021] Furthermore, the detection probability is:

[0022]

[0023] Among them, P d(i,j) represents the detection probability of the jth primary user channel by the i-th secondary user; Q represents the right tail function of the standard normal distribution; ε i represents the threshold value of the i-th user; represents the variance of the noise; γ i,j Represents the average signal-to-noise ratio received by the secondary user.

[0024] Furthermore, in S3, in the information transmission stage of cooperative spectrum sensing, clustering constraints are set, and according to the constraints and the detection probability matrix, the primary user and the secondary user are clustered using a bipartite graph to obtain multiple clusters. The specific process is as follows:

[0025] S31, setting clustering restriction conditions;

[0026] S32, taking the secondary user and primary user channels as two point sets of a bipartite graph, and the detection probability matrix as a set of edges in the bipartite graph, and decomposing the bipartite graph into multiple complete bipartite graphs according to the clustering restriction condition, each complete bipartite graph represents a cluster, and the two point sets of each complete bipartite graph correspond to the secondary user and primary user channels in the cluster respectively;

[0027] The secondary users and primary users in each cluster are removed from the detection probability matrix to obtain a new detection probability matrix. The above operation is repeated according to the new detection probability matrix until the detection probability of each primary user channel is greater than the corresponding detection probability threshold or the maximum number of iterations is reached, and the clustering is completed to obtain multiple clusters.

[0028] Furthermore, the restriction conditions for clustering in S31 are:

[0029] a) Each secondary user is divided into at most three different clusters, and each primary user channel can only be detected by one cluster;

[0030] b) using an OR fusion rule to perform fusion detection on the spectrum sensing results of all secondary users corresponding to a primary user channel to obtain the detection probability of the primary user channel, and ensuring that the detection probability of the primary user channel is greater than a custom detection probability threshold;

[0031] c) Under the condition of b), the sum of the detection probabilities of all primary user channels is maximized.

[0032] A multi-channel cooperative spectrum sensing system comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, any step of a multi-channel cooperative spectrum sensing method is implemented.

[0033] A computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements any step of a multi-channel cooperative spectrum sensing method.

[0034] Beneficial effects:

[0035] The present invention defines that the main user and the secondary user in the cognitive wireless network obey the homogeneous Poisson point process distribution respectively in the cognitive wireless network, and the two homogeneous Poisson point processes are independent of each other. In the local perception process of collaborative spectrum sensing, each secondary user performs energy detection on each main user channel in turn, that is, the main user signal received by the secondary user is processed in turn through a bandpass filter, a square law element, an analog-to-digital converter, an integrator or an accumulator (discrete signal) to obtain an energy detection statistic, and the energy detection statistic is compared with a self-defined threshold value. If the energy detection statistic is greater than the threshold value, it indicates that the main user signal exists. The detection probability and the false alarm probability are calculated according to the energy detection statistic to obtain a detection probability matrix. If the energy detection statistic is less than the threshold value, it indicates that the main user signal does not exist. The secondary user and the main user channels are clustered, and the restriction conditions during clustering are set. The preliminary clustering is completed according to the restriction conditions and the bipartite graph. After a cluster is divided, the rows corresponding to the secondary users and the columns corresponding to the primary users in the cluster are removed from the detection probability matrix, and the detection probability matrix with the corresponding rows and columns removed is used as the detection probability matrix for the next round of iteration, until the collaborative detection probability of all secondary users of the primary user channel is greater than the custom detection probability threshold or the upper limit of the number of iterations is met, and the clustering is completed. The secondary user with the largest receiving signal-to-noise ratio in the cluster is elected as the cluster head, which is responsible for collecting and fusing the spectrum perception information of other secondary users in the cluster, and interacting with other cluster heads with the judgment results. The spectrum perception results of all secondary users of each primary user channel are fused and detected using the OR fusion rule to obtain the spectrum perception results of each primary user channel and determine the occupancy status of each primary user channel.

[0036] The present invention utilizes the similarity of the spectrum sensing results of secondary users for clustering, divides the secondary users with large similarity of spectrum sensing results into a cluster, and then allocates the main user channels of these secondary users with good detection performance to the cluster for detection, and finally merges the sensing results of all clusters to obtain the occupancy status of all main user channels. This method does not require knowing the geographical location of each secondary user in advance, and this clustering method based on spectrum sensing results can take into account the influence of various fading on the secondary users, making the local sensing results more reliable, thereby greatly improving the accuracy and real-time performance of the collaborative spectrum sensing results. This method does not assume that the sensing results of the secondary users are exactly the same, but assumes that the sensing results between the secondary users are similar, which is in line with the actual situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a schematic diagram of a two-layer heterogeneous network of cognitive radio;

[0038] Figure 2 is a flow chart of the energy detection algorithm;

[0039] Figure 3 It is a schematic diagram of a bipartite graph;

[0040] Figure 4 is a schematic diagram of a complete bipartite graph;

[0041] Figure 5 is a bipartite graph corresponding to the clustering result of the present invention; DETAILED DESCRIPTION

[0042] Specific implementation method 1: Combination Figure 1-Figure 5 This embodiment describes a multi-channel collaborative spectrum sensing method, which includes the following steps:

[0043] S1. Acquire a cognitive wireless network, obtain primary user information and secondary user information in the cognitive wireless network, wherein the primary user and the secondary user obey homogeneous Poisson point process distribution in the cognitive wireless network, and the two homogeneous Poisson point processes are independent of each other. The information includes quantity and location, etc.

[0044] In an actual radio environment, the spatial distribution of each node (secondary user and primary user) in a cognitive wireless network is random. Therefore, in order to be more in line with the actual communication scenario, the present invention introduces random geometry theory into the modeling of cognitive wireless networks, and uses random point processes to describe the random distribution of each node (secondary user and primary user) in cognitive wireless networks. Among various point process models, the most widely used one is the Poisson Point Process (PPP). Therefore, the present invention uses the Poisson Point Process to describe the spatial distribution of each node (secondary user and primary user).

[0045] When a point process in a certain space in a cognitive wireless network satisfies the following two conditions at the same time, it is called a homogeneous Poisson point process (HPPP).

[0046] Condition 1: For any bounded region D in a cognitive wireless network, if the number of nodes N(D) in the region follows a Poisson distribution, then the mean is λv d (D), where λ is the Poisson strength, which needs to be customized, λ>0,v d (D) is the area of ​​region D, expressed as:

[0047]

[0048] Wherein, P(N(D)=k) represents the probability that the number of points in region D is k.

[0049] Condition 2: There are disjoint bounded regions D in cognitive wireless networks 1 ,D 2 , ..., D n ,The total number of nodes in these regions are independent of each other.

[0050] Based on the above conditions, the homogeneous Poisson point process has the following properties:

[0051] 1. The nodes of the homogeneous Poisson point process are distributed independently in space, and there is no correlation between the nodes.

[0052] 2. In the same Poisson point process, the number of nodes per unit area in different regions is the same.

[0053] 3. The superposition of two homogeneous Poisson point processes is still a homogeneous Poisson point process, and its Poisson intensity is the sum of the two.

[0054] 4. When some nodes in the homogeneous Poisson point process are selected with independent probability p and other points are eliminated with probability 1-p, two independent homogeneous Poisson point processes can be obtained, with Poisson intensities pλ and (1-p)λ respectively.

[0055] Since cognitive wireless networks contain two different types of communication nodes: primary user PU and secondary user SU, the present invention models the primary user PU and secondary user SU as independent Poisson point processes, and finally forms a two-layer heterogeneous network. Assuming that the primary user node is a homogeneous Poisson point process with a Poisson strength of 5, and the secondary user node is a homogeneous Poisson point process with a Poisson strength of 15, the corresponding two-layer heterogeneous network is as follows: Figure 1 As shown, the asterisk represents the primary user PU, and the hollow origin represents the secondary user SU. Assuming that each primary user PU occupies a channel to transmit information, and all channels are currently occupied by primary users, the goal of the present invention is to mobilize the available secondary users to quickly and accurately perceive the existence of primary user signals on all channels. To solve this problem, all secondary users can detect all primary user signals in turn, and finally fuse the perception results of the secondary users according to certain fusion rules to obtain the final judgment result. However, this method will bring a large perception delay, so it is necessary to assign different perception tasks to different secondary users to improve the real-time performance of perception. Therefore, it is particularly important to design a reasonable collaborative spectrum sensing strategy.

[0056] S2. In the local sensing stage of cooperative spectrum sensing, each secondary user performs energy detection on each primary user channel in turn to obtain a detection probability matrix. The specific process is as follows:

[0057] First, in the local sensing process of the cooperative spectrum sensing technology, a low-complexity energy detection method is used to enable each secondary user to perform energy detection on each primary user channel in turn. The working principle block diagram of the energy detection method is as follows: Figure 2 As shown. Figure 2It can be obtained that the secondary user processes the received primary user signal through a bandpass filter, a square law element, an analog-to-digital converter, an integrator or an accumulator (discrete signal) in sequence to obtain an energy detection statistic, and compares the energy detection statistic with a user-defined threshold value. If the energy detection statistic is greater than the threshold value, it indicates that the primary user signal exists. The detection probability and false alarm probability of the primary user are calculated according to the energy detection statistic to obtain a detection probability matrix. If the energy detection statistic is less than the threshold value, it indicates that the primary user signal does not exist.

[0058] The bandpass filter is used to remove out-of-band noise. Since the primary user signal occupies a certain bandwidth, it will be affected by noise during the transmission process, resulting in the frequency domain of the primary user signal at the receiving end being widened. Therefore, a bandpass filter is usually placed at the receiving end to filter out the signal outside the bandwidth of the primary user signal, that is, the out-of-band noise, so that the received signal is closer to the primary user signal. Generally, in practical applications, integrators or accumulators (discrete signals) are used for summation, and the summation process will adopt the processing method of discrete signals. Therefore, an analog-to-digital converter (A / D) is added between the square law element and the integrator or accumulator, and the signal is Nyquist sampled to obtain the energy detection statistic. Finally, the energy detection statistic is compared with the threshold value to obtain the final judgment result. Then the binary perception model of the i-th secondary user to the j-th primary user channel is:

[0059]

[0060] Where y(n) is the signal after y(t) is sampled. y(t) refers to the signal received by the secondary user, which is composed of the primary user signal and noise. i,j (n) represents the binary perception model of the i-th secondary user to the j-th primary user channel, s i,j (n) represents the signal of the jth primary user that the i-th secondary user can receive, and s is defined as i,j The mean of (n) is 0 and the variance is w i,j (n) is the noise, and w is defined i,j The mean of (n) is 0 and the variance is Then the energy detection statistic Y can be obtained as:

[0061]

[0062] Where Y i,j represents the energy detection statistic of the i-th secondary user to the j-th primary user signal; f s is the sampling frequency, τ is the sampling time. Then the detection probability P of the energy detection algorithm d and false alarm probability P f They are:

[0063]

[0064]

[0065] Where P d(i,j) represents the detection probability of the i-th secondary user on the j-th primary user channel; ε i is the threshold value of the ith user, γ i,j is the average signal-to-noise ratio received by the secondary user, Q is the Q function, which represents the right tail function of the standard normal distribution. d The detection probability matrix is ​​composed of the detection probability matrix, which is the detection probability of each secondary user for each primary user channel, and is one-to-one. The present invention assumes that the signal-to-noise ratio is a known condition, Q(x) and γ i,j They are:

[0066]

[0067] In the formula, t has no special meaning and is just an integral variable.

[0068]

[0069] S3. In the information transmission stage of cooperative spectrum sensing, set the restriction conditions for clustering. According to the restriction conditions and the detection probability matrix, use the bipartite graph to cluster the primary users and secondary users to obtain multiple clusters. The specific process is as follows:

[0070] After solving the problem of local perception of secondary users, the next step is to consider how to reasonably cluster the secondary users to ensure accurate and rapid acquisition of the occupancy status of all primary user channels. The following is a detailed analysis of the clustering problem. First of all, it should be made clear that the secondary users assigned to the same cluster are responsible for detecting the channel status of the primary users assigned to this cluster, and it is expected that the final clustering result can maximize the overall detection accuracy of the system.

[0071] S31. Set the restriction conditions for clustering:

[0072] a) In order to ensure the perception performance of the primary user channel, each secondary user is divided into at most three different clusters, and each primary user channel can only be detected by one cluster.

[0073] b) Use the OR fusion rule to fuse the spectrum sensing results of all secondary users corresponding to a primary user channel to obtain the detection probability P of this primary user channel. d , ensuring the detection probability P of this primary user channel d Greater than the custom detection probability threshold P thThe detection probability threshold is given, and when the cooperative detection probability of all secondary users detecting the primary user channel is greater than the detection threshold, it is considered that these secondary users can form a cluster to detect the primary user channel, otherwise, they cannot form a cluster.

[0074] Traditional collaborative spectrum sensing strategies allocate sensing tasks based on a single secondary user. When the secondary user transmits local sensing information to the fusion center, it will bring a large communication overhead. Therefore, the existing technology introduces clustering technology into collaborative spectrum sensing technology, that is, to divide the secondary users into different clusters, allocate sensing tasks based on clusters, and finally merge the sensing results of each cluster. However, most of them divide clusters based on the location information of the secondary users. This method is not only low in complexity, but also difficult to obtain the location information of the secondary users. In addition, due to the different geographical locations of different secondary users, the results of sensing the same frequency band are also different. In particular, when the secondary user is blocked by obstacles, it may not detect the existence of signals in this frequency band. When there is no blockage, it may sense the existence of signals. The perception accuracy of this clustering method will be greatly reduced, and the spectrum sensing results are subjectively judged by each secondary user. However, whether there is a signal in this frequency band is certain and objective, and it will not exist just because the secondary user does not sense it.

[0075] Since the clustering-based collaborative spectrum sensing technology can not only reduce the communication overhead of the system, but also make the topology of the cognitive wireless network easy to manage. Therefore, the present invention designs a clustering method based on spectrum sensing results, which divides clusters by using the similarity of the secondary user spectrum sensing results, allowing multiple secondary users in a cluster to sense a primary user channel, which not only overcomes the disadvantage that the secondary user position is difficult to obtain, but also effectively resists the influence of shadow effects, etc., and can greatly improve the accuracy of spectrum sensing. Taking the jth primary user channel as an example, its collaborative detection probability and collaborative false alarm probability are:

[0076]

[0077]

[0078] c) Under the condition of b), the sum of the detection probabilities of all primary user channels is maximized as much as possible.

[0079] In summary, this constraint can be described in mathematical language as:

[0080]

[0081] In the formula, θ k represents the sum of the cooperative detection probabilities of all primary user channels contained in the kth cluster; S krepresents the secondary user number vector contained in the kth cluster. Since a cluster may contain multiple secondary users, this cluster can be represented by a vector composed of the sequence numbers of all secondary users in the cluster. For example, the kth cluster contains the 1st and 3rd secondary users, then S k =(1,3). C k represents the sequence number vector of the primary user channels to be detected in the kth cluster. For example, if the kth cluster wants to detect the second and fourth primary user channels, then C k =(2,4);X S(N×K) and X P(M×K) They represent the secondary user allocation matrix and the primary user channel allocation matrix respectively, where N is the number of secondary users in the cognitive network, M is the number of primary user channels to be detected, K is the total number of clusters, and X S(N×K) and X P(M×K) The meaning of each element is shown in the following formula.

[0082]

[0083]

[0084] By analyzing the above mathematical model (Formula (10)), it can be obtained that the optimization goal of the multi-channel cooperative spectrum sensing of the present invention is to maximize the sum of the detection probabilities of all primary user channels. The first constraint in Formula (10) ensures that the detection probability of each primary user channel is greater than the set detection probability threshold, the second constraint constrains a secondary user to be assigned to at most 3 clusters, and the third constraint constrains a primary user channel to be assigned to only one cluster for detection.

[0085] S32, taking the secondary user and primary user channels as two point sets of a bipartite graph, and the detection probability matrix as a set of edges in the bipartite graph, and decomposing the bipartite graph into multiple complete bipartite graphs according to the clustering restriction condition, each complete bipartite graph represents a cluster, and the two point sets of each complete bipartite graph correspond to the secondary user and primary user channels in the cluster respectively;

[0086] Remove the secondary users and primary users in each cluster from the detection probability matrix to obtain a new detection probability matrix. Repeat the above operation according to the new detection probability matrix until the detection probability of each primary user channel is greater than the corresponding detection probability threshold or the maximum number of iterations is reached, and clustering is completed to obtain multiple clusters. The maximum number of iterations can be customized.

[0087] The present invention uses a bipartite graph to solve formula (10). A bipartite graph is also called a bipartite graph, which is a special model in graph theory. Let G = (V, E) be an undirected graph. If the vertex V can be divided into two non-intersecting subsets (A, B), and the two vertices i and j associated with each edge E (i, j) in the graph belong to these two different vertex sets (i∈A, i∈B), then the graph G is called a bipartite graph. In other words, if the vertices of a graph can be divided into two non-intersecting subsets, and each edge in the graph connects the vertices in the two sets, then the graph is a bipartite graph, such as Figure 3 As shown. A bipartite graph only requires two sets of points and one set of edges to be determined.

[0088] A complete bipartite graph is a special bipartite graph that adds a restriction to the bipartite graph. That is, after the points in the graph are divided into two non-intersecting point sets, all vertices in the first set are connected to all vertices in the second set. Therefore, a complete bipartite graph only needs two point sets to be determined, such as Figure 4 shown.

[0089] Corresponding to the mathematical model (Formula (10)), the two point sets of the bipartite graph represent the secondary user and primary user channels respectively, and the corresponding detection probability matrix is ​​the weight of the edge connecting the two. According to the principle of maximizing the sum of the detection probabilities of the primary user channels, the clustering problem is transformed into the problem of decomposing a weighted bipartite graph into multiple complete bipartite graphs. Each complete bipartite graph represents a clustering result, and the two point sets of each complete bipartite graph correspond to the secondary users in the cluster and the primary user channels to be detected. After a cluster is divided, the rows corresponding to the secondary users and the columns corresponding to the primary users in the cluster are removed from the detection probability matrix to obtain a new detection probability matrix. The bipartite graph is repeatedly divided or assembled according to the new detection probability matrix, and the new detection probability matrix is ​​used as the detection probability matrix for the next round of iteration until the detection probability of each primary user channel is greater than the corresponding detection probability threshold or the maximum number of iterations is reached, the clustering is completed, and the clustering result is output.

[0090] The purpose of decomposition is to maximize the sum of the detection probabilities of all primary user channels.

[0091] S4. In the data fusion stage of collaborative spectrum sensing, after the clustering is completed, the secondary user with the largest average signal-to-noise ratio in each cluster is selected as the cluster head. The cluster head is responsible for collecting the spectrum sensing information of other secondary users in the cluster, and using the OR fusion rule to fuse the spectrum sensing information of other secondary users. According to the fused spectrum sensing information, it is determined whether there is a primary user signal in the primary user channel corresponding to the cluster, and the presence of the primary user signal is interactively fused with other cluster heads. After the interaction, the occupancy status of all primary user channels can be obtained. The occupancy status indicates whether there is a signal in the primary user.

[0092] S5. Obtain the cognitive wireless network to be identified, and repeat S1-S4 to obtain the occupancy status of all primary user channels.

[0093] Specific implementation method 2: Combination Figure 1-Figure 5 To illustrate this embodiment, a multi-channel collaborative spectrum sensing system described in this embodiment includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any step of a multi-channel collaborative spectrum sensing method.

[0094] Specific implementation method three: Combination Figure 1-Figure 5 To illustrate this embodiment, a computer-readable storage medium is described in this embodiment, and the computer-readable storage medium stores a computer program. When the computer program is executed by a processor, any step of a multi-channel collaborative spectrum sensing method is implemented.

[0095] Example 1

[0096]

[0097] Example 2

[0098] The present invention is simulated in the following simulation environment: the cognitive radio network contains 12 secondary users and 6 primary users, and the primary users and secondary users obey the Poisson point process distribution in space, and the two Poisson point processes are independent of each other; each primary user occupies a channel to transmit information, and the transmission frequencies are 50, 70, 90, 110, 130 and 150 MHz, respectively, and the transmission power is 40 kW. The product of sampling frequency and sampling time is f s τ=20;receiver noise power is set to -53dBm, detection threshold is -48dBm. The distance matrix D from each secondary user to the primary user is (in km):

[0099]

[0100] And considering only shadow fading and path loss, the received power can be expressed as:

[0101] P r =P t L 0 L 1 (1)

[0102]

[0103] Among them, L 0 represents the path loss, L 1 represents the loss caused by shadow fading, λ cis the signal wavelength, d is the distance between the receiver and the transmitter, G t and G r are the gains of the transmitting antenna and the receiving antenna respectively.

[0104] Here, the received signal-to-noise ratio of each secondary user is obtained according to the strength of the received signal, and then the corresponding detection probability is calculated according to the detection probability calculation formula of energy detection, so as to obtain the corresponding detection probability matrix:

[0105]

[0106] Among them, p i,j Represents the detection probability of the i-th node for the j-th channel. According to the detection probability matrix combined with the bipartite graph model, and set the detection probability threshold Q th is 0.98, and the simulation results of clustering can be obtained as follows Figure 5 As shown. By analyzing the simulation results, it can be obtained that the clustering result has the constraint that a primary user channel can only be assigned to one cluster, a secondary user is not assigned to a maximum of three clusters, and secondary users with poor channel perception performance for all primary users do not participate in this collaborative perception, which is in line with the actual situation.

Claims

1. A multi-channel collaborative spectrum sensing method, Features: It includes the following steps: S1. Acquire a cognitive wireless network, obtain primary user information and secondary user information in the cognitive wireless network, wherein the primary user and the secondary user obey homogeneous Poisson point process distributions in the cognitive wireless network, and the two homogeneous Poisson point processes are independent of each other; S2. In the local sensing stage of collaborative spectrum sensing, each secondary user performs energy detection on each primary user channel in turn to obtain a detection probability matrix; S3. In the information transmission stage of cooperative spectrum sensing, set the restriction conditions for clustering. According to the restriction conditions and the detection probability matrix, use the bipartite graph to cluster the primary users and secondary users to obtain multiple clusters. The specific process is as follows: S31. Set the restriction conditions for clustering: a) Each secondary user is divided into at most three different clusters, and each primary user channel can only be detected by one cluster; b) using an OR fusion rule to perform fusion detection on the spectrum sensing results of all secondary users corresponding to a primary user channel to obtain the detection probability of the primary user channel, and ensuring that the detection probability of the primary user channel is greater than a custom detection probability threshold; c) Under the condition of b), the sum of the detection probabilities of all primary user channels is maximized; S32, taking the secondary user and primary user channels as two point sets of a bipartite graph, and the detection probability matrix as a set of edges in the bipartite graph, and decomposing the bipartite graph into multiple complete bipartite graphs according to the clustering restriction condition, each complete bipartite graph represents a cluster, and the two point sets of each complete bipartite graph correspond to the secondary user and primary user channels in the cluster respectively; The secondary users and primary users in each cluster are removed from the detection probability matrix to obtain a new detection probability matrix. The above operation is repeated according to the new detection probability matrix until the detection probability of each primary user channel is greater than the corresponding detection probability threshold or the maximum number of iterations is reached, and clustering is completed to obtain multiple clusters. S4. In the data fusion stage of collaborative spectrum sensing, the secondary user with the largest average signal-to-noise ratio of the primary user signal received in each cluster is selected as the cluster head. The cluster head collects the spectrum sensing information of other secondary users in the cluster, and fuses the spectrum sensing information of other secondary users using the OR fusion rule. It is determined whether there is a primary user signal in the primary user channel corresponding to the cluster based on the fused spectrum sensing information, and the existence of the primary user signal is interactively fused with other cluster heads to obtain the occupancy status of all primary user channels. S5. Obtain the cognitive wireless network to be identified, and repeat S1-S4 to obtain the occupancy status of all primary user channels.

2. According to the multi-channel collaborative spectrum sensing method described in claim 1, Features: The primary user information in S1 includes the number of primary users and the position of each primary user in the cognitive wireless network, and the secondary user information includes the number of secondary users and the position of each secondary user in the cognitive wireless network.

3. According to the multi-channel collaborative spectrum sensing method described in claim 2, Features: The homogeneous Poisson point process in S1 is: When a point process in a certain space in a cognitive wireless network satisfies the following two conditions at the same time, it is called a homogeneous Poisson point process; Condition 1: For any bounded area D in a cognitive wireless network, the number of nodes N(D) in the area obeys a Poisson distribution; Condition 2: There are disjoint bounded regions D in cognitive wireless networks 1 ,D 2 , ..., D n , the total number of nodes in the regions are independent of each other.

4. According to the multi-channel collaborative spectrum sensing method described in claim 3, Features: In the local sensing process of cooperative spectrum sensing in S2, each secondary user performs energy detection on each primary user channel in turn to obtain a detection probability matrix. The specific process is as follows: The secondary user processes the received primary user signal through a bandpass filter, a square-law element, an analog-to-digital converter, an integrator or an accumulator in sequence to obtain an energy detection statistic, and compares the energy detection statistic with a custom threshold value. If the energy detection statistic is greater than the threshold value, it indicates that the primary user signal exists, and the detection probability of the primary user is calculated to obtain a detection probability matrix. The detection probability matrix includes the detection probability of each secondary user for each primary user channel; if the energy detection statistic is less than the threshold value, it indicates that the primary user signal does not exist.

5. According to the multi-channel collaborative spectrum sensing method described in claim 4, Features: The energy detection statistics are: Among them, Y i,j represents the energy detection statistic of the i-th secondary user to the j-th primary user signal; f s is the sampling frequency; τ is the sampling time; y i,j (n) represents the binary perception model of the i-th secondary user to the j-th primary user channel.

6. According to the multi-channel collaborative spectrum sensing method described in claim 5, Features: The detection probability is: Among them, P d(i,j) represents the detection probability of the jth primary user channel by the i-th secondary user; Q represents the right tail function of the standard normal distribution; ε i represents the threshold value of the i-th user; represents the variance of the noise; γ i,j Represents the average signal-to-noise ratio received by the secondary user.

7. A multi-channel cooperative spectrum sensing system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, Features: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

8. A computer-readable storage medium storing a computer program. Features: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images