A spectrum cognitive collaborative anti-interference method, device, storage medium and application

Abstract

The present application discloses a spectrum cognitive collaborative anti-interference method, which includes: S1 all cognitive users establish an initialized spectrum perception neural network; S2 cognitive users perform full-band perception within the frequency band, and at the same time label all cognitive users according to the perceived spectrum data and store them in a data pool; randomly batch sample spectrum data from the data pool to train the spectrum perception neural network; S3 repeat step S2 to a preset number of cycles; S4 cognitive users select channels based on the greedy criterion to obtain an average frequency utilization success rate; S5 cognitive users obtain a loss function based on the average frequency utilization success rate, update the initially trained spectrum perception neural network, and obtain the cognitive user's anti-interference strategy based on the trained spectrum perception neural network. The present invention also discloses a spectrum cognitive collaborative anti-interference device and storage medium. The present invention adopts an intelligent perception method, and can obtain an efficient and reliable collaborative anti-interference communication strategy through independent learning.

Description

Technical Field

[0001] The present application relates to the field of wireless anti-interference technology, and more specifically, to a spectrum-aware collaborative anti-interference method, device, storage medium, and application. Background Art

[0002] Wireless technology is making people's lives more convenient and intelligent. However, the openness of the electromagnetic spectrum is a double-edged sword, exposing the security of wireless communication systems to many threats. Among them, wireless jammers can easily paralyze wireless networks by releasing interference.

[0003] In a multi-user wireless network where surrounding frequency-using devices and malicious interference devices coexist, multiple cognitive users are simultaneously faced with co-frequency interference between internal users, unintentional interference from external frequency-using devices, and malicious interference from jammers, which significantly increases the difficulty of spectrum access to wireless communication networks.

[0004] Existing learning-based anti-interference communication methods adjust anti-interference strategies based on the communication effects obtained after actual spectrum access, and rely on information interaction between users to achieve collaboration. Such methods have low learning efficiency, poor online learning effects, and are prone to causing unintentional interference to other users' communications. They are also difficult to be effective in environments with severe interference. Summary of the Invention

[0005] In response to at least one defect or improvement need in the prior art, the present invention provides a spectrum-aware collaborative anti-interference method, device, storage medium and application to solve at least one of the above-mentioned technical problems.

[0006] To achieve the above objectives, according to a first aspect of the present invention, a spectrum-aware collaborative anti-interference method is provided, which utilizes a trained spectrum-aware neural network to obtain a channel selection strategy and implement anti-interference communication. The method includes:

[0007] S1 establishes an initialized spectrum sensing neural network for all cognitive users;

[0008] All cognitive users in S2 perform full-band perception within the frequency band and obtain spectrum data including spectrum occupancy status of any time slot; all cognitive users perform channel access and communication based on preset frequency hopping rules, and at the same time, all spectrum data perceived by cognitive users are labeled and stored in the data pool

[0009] Randomly batch sample spectrum data from the data pool to train the spectrum perception neural network;

[0010] S3 repeats step S2 for a preset number of cycles to obtain a spectrum sensing neural network after preliminary training;

[0011] S4: All cognitive users perform full-band sensing within the frequency band to obtain the spectrum occupancy status of any time slot. All cognitive users select channels based on the greedy criterion and communicate on their selected channels to obtain the average frequency utilization success rate.

[0012] S5. All cognitive users determine a loss function for training the preliminarily trained spectrum sensing neural network based on the average frequency utilization success rate, calculate the gradient of the loss function, and update the preliminarily trained spectrum sensing neural network based on the gradient; and obtain the anti-interference strategy of the cognitive user based on the trained and updated spectrum sensing neural network.

[0013] Furthermore, the above-mentioned spectrum cognitive collaborative anti-interference method also includes: the spectrum perception neural network includes a pooling layer, a convolution layer and a fully connected layer, wherein: the pooling layer is used to reduce redundant information in the spectrum data, reduce the matrix dimension and thus reduce the amount of calculation; the convolution layer is used to extract the characteristics of signal occupancy in the spectrum data; and the fully connected layer is used to fit the frequency utilization strategy.

[0014] Furthermore, in the spectrum cognitive collaborative anti-interference method, in step S2, all cognitive users label the spectrum data they perceive and store it in a data pool, and randomly batch sample spectrum data from the data pool to train the spectrum perception neural network, specifically including:

[0015] Label the perceived spectrum data and obtain the labeled spectrum data in is the spectrum occupancy status of the t-th time slot, is an M-dimensional vector, In the tth time slot, cognitive user n perceives the occupancy state of the Mth channel; the output of the spectrum sensing neural network is Define the loss function as The gradient of the loss function is calculated, and a gradient-based optimization method is used to train the spectrum sensing neural network.

[0016] Furthermore, in the above spectrum cognitive collaborative anti-interference method, in step S4, all cognitive users select channels based on the greedy criterion and communicate on the channels selected by them to obtain an average frequency utilization success rate, specifically including: all cognitive users select channels based on the greedy criterion, that is, input the spectrum data of the kth time slot into the spectrum sensing neural network. Spectrum sensing neural network output Select the action corresponding to the maximum value as the result of the decision, that is

[0017] Among them, the frequency success rate is in

[0018]

[0019] Where θ∈(0.9,1) is a constant, Indicates whether the frequency usage of the kth time slot is successful. If the frequency usage is successful, If the frequency fails,

[0020] Furthermore, in the spectrum cognitive collaborative anti-interference method, in step S5, all cognitive users determine a loss function for training the preliminarily trained spectrum sensing neural network based on the average frequency utilization success rate, calculate the gradient of the loss function, and update the preliminarily trained spectrum sensing neural network based on the gradient, specifically including:

[0021] For cognitive user n, define the loss function The gradient of the loss function is calculated and the neural network is trained using a gradient-based optimization method.

[0022] Furthermore, the spectrum-aware collaborative anti-interference method further includes:

[0023] Before step S1, a cognitive user is randomly selected as the control user;

[0024] After step S5, if the frequency utilization success rate of the control user is lower than the first threshold, all cognitive users return to step S2 in a preset time slot.

[0025] According to a second aspect of the present invention, a multi-user anti-interference frequency game method based on the above-mentioned spectrum-aware collaborative anti-interference method is also provided, which is characterized by comprising:

[0026] A multi-user anti-interference frequency game model is established, which includes traditional communication equipment, surrounding frequency-using equipment, malicious interference equipment and multiple cognitive users. The frequency-using strategy of the traditional communication equipment is π C , the frequency usage strategy of the surrounding frequency-using devices is π A , the frequency strategy of the malicious interference device is π J , the frequency strategy of cognitive user n is π n , the cognitive user n is one of a plurality of cognitive users 1, 2, ..., N;

[0027] The model is G = {N, S, A n , r n}, where N is the total number of cognitive users, S is the set of spectrum states obtained by all cognitive users at a certain moment, and A n is the action set of cognitive user n, r n is the reward function of cognitive user n;

[0028] The average frequency success rate of cognitive user n in the tth time slot As the reward of cognitive user n in time slot t The average frequency success rate for

[0029]

[0030] Where θ∈(0.9,1) is a constant, Indicates whether the frequency usage of the tth time slot is successful. If the frequency usage is successful, If the frequency fails,

[0031] The cognitive user n obtains a frequency utilization strategy based on the spectrum sensing neural network;

[0032] For cognitive user n, the multi-user anti-interference frequency game model is constructed to meet the following conditions:

[0033] maxη n

[0034] st(π n ∩π c )=0

[0035] The cognitive user n selects a frequency utilization strategy that meets the above conditions based on the multi-user anti-interference frequency utilization game model.

[0036] Furthermore, the above multi-user anti-interference frequency game method, Indicates whether the frequency usage of the tth time slot is successful, including:

[0037]

[0038] Where δ(X) is the indicator function. If the X condition is met, then δ(X) = 1. If the X condition is not met, then δ(X) = 0. is the signal-to-interference-and-noise ratio received by cognitive user n, which is related to the frequency utilization strategies adopted by frequency-using devices and malicious interference devices around the tth time slot, λ th is the communication service quality requirement threshold.

[0039] According to the third aspect of the present invention, a spectrum-aware collaborative anti-interference device is also provided, which includes at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit performs the steps of any one of the above methods.

[0040] According to a fourth aspect of the present invention, a storage medium is provided, which stores a computer program executable by an access authentication device. When the computer program runs on the access authentication device, the access authentication device executes the steps of any one of the above methods.

[0041] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0042] The present invention provides a spectrum-aware collaborative anti-interference method, which adopts a deep learning method for intelligent spectrum perception, utilizes the state information obtained by full-band perception, and adopts a supervised learning method to improve learning efficiency. During the learning process, no information interaction between users is required, and multi-user collaborative anti-interference strategies can be quickly learned, effectively improving the collaborative anti-interference capability under complex spectrum conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 A schematic diagram of a spectrum-aware collaborative anti-interference method provided in an embodiment of the present application;

[0045] Figure 2 A schematic diagram of a communication environment provided in an embodiment of the present application;

[0046] Figure 3 This is a simulation diagram showing how the average frequency success rate of the algorithm proposed in the embodiment of the present application changes with the number of iterations. DETAILED DESCRIPTION

[0047] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0048] The terms "first," "second," "third," and the like in the specification and claims of this application and the accompanying drawings are used to distinguish between different objects, not to describe a particular order. Furthermore, the terms "including," "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or elements is not limited to the listed steps or elements, but may optionally include steps or elements not listed, or may optionally include other steps or elements inherent to the process, method, product, or apparatus.

[0049] The following describes the use scenario of the spectrum cognitive collaborative anti-interference method involved in this application: In a distributed multi-user wireless communication network, multiple cognitive users share spectrum resources for wireless communication. Each cognitive user has the ability to independently perceive the spectrum, learn frequency utilization strategies, and make frequency utilization decisions. It can optimize its own frequency utilization strategy in real time according to the spectrum environment status to improve the communication success rate. Outside the wireless network, there are traditional communication devices (such as base stations and mobile phones that communicate on authorized frequency bands), surrounding frequency-using devices (such as radars), and malicious interference devices, which affect the communications of wireless network users.

[0050] The communication parameters and frequency usage strategies of traditional communication equipment are preset and fixed. Such equipment does not have the ability to learn and make decisions independently, and its communication parameters and frequency usage rules are usually public. Cognitive users need to actively avoid the signals of traditional communication equipment when communicating to avoid interfering with traditional communication systems; the frequency bands used by surrounding frequency-using equipment overlap with those used by the communication network, which will cause unexpected interference to the communication network. Cognitive users need to learn the frequency usage patterns of such frequency-using equipment to avoid interference; malicious interference equipment attacks the information transmission of wireless network users by sending dynamic malicious interference signals. It is more dynamic and confrontational, and has a more serious impact on the communication of cognitive users.

[0051] In this heterogeneous, dynamic, and adversarial complex spectrum environment, distributed multi-cognitive users need to jointly consider various external interferences to avoid them. At the same time, they need to coordinate the frequency usage among cognitive users to avoid mutual interference between users. At this time, a multi-cognitive user anti-interference method is needed to help cognitive users formulate frequency usage strategies.

[0052] like Figure 1 As shown in one embodiment, it is a process diagram of a spectrum-aware collaborative anti-interference method provided by the present invention, which implements anti-interference communication by obtaining a channel selection strategy using a trained spectrum-aware neural network, and provides a spectrum-aware collaborative anti-interference method, including the following steps:

[0053] S1 establishes an initialized spectrum sensing neural network for all cognitive users.

[0054] Specifically, all cognitive users can share a spectrum sensing neural network. First, the spectrum sensing neural network is initialized. The initialization method can be selected according to the actual situation. Commonly used initialization methods include all-zero or equal value initialization, normal initialization, uniform initialization, Xavier, He initialization, orthogonal and other methods. By initializing the spectrum sensing neural network, the problem of gradient disappearance or gradient explosion caused by incorrect initialization weights can be effectively avoided, thereby avoiding negative impacts on the training process.

[0055] In S2, all cognitive users perform full-band perception within the frequency band and obtain spectrum data including spectrum occupancy status for any time slot; all cognitive users perform channel access and communication based on preset frequency hopping rules, and at the same time, the spectrum data perceived by all cognitive users are labeled and stored in a data pool; spectrum data is randomly batch sampled from the data pool to train the spectrum perception neural network.

[0056] Specifically, each cognitive user has independent spectrum sensing capabilities and can sense the spectrum occupancy status. All cognitive users perform full-band sensing within the frequency band at the tth time slot and can obtain the spectrum sensing status including the tth moment. Spectrum data, where n = 1, ..., N, represents one of the N cognitive users, and the spectrum occupancy status Is a Φ-dimensional vector, where Φ is the number of spectrum sensing sampling points. The larger the Φ value, the The more detailed the spectrum state is described, the larger the spectrum data dimension is; The value is the energy value of the sampling point. The spectrum data perceived by all cognitive users in T time slots Combined into a two-dimensional spectrum data matrix.

[0057] The preset frequency hopping rule is that all cognitive users access the spectrum according to a pre-set frequency hopping sequence. The frequency hopping sequences between cognitive users are orthogonal to each other, and the frequency hopping sequences of all cognitive users are orthogonal to the frequencies used by traditional devices.

[0058] All cognitive users can calculate the signal power in each channel based on the perceived spectrum data, determine whether each channel is occupied, and use the channel occupancy status as a label to label the perceived spectrum data and store it in the data pool.

[0059] Spectrum data is randomly sampled in batches from the data pool, and the sampled spectrum data is used to train the spectrum perception neural network.

[0060] S3 repeats step S2 for a preset number of cycles to obtain a spectrum sensing neural network after preliminary training.

[0061] Specifically, the number of cycles can be preset according to actual needs, so that the spectrum sensing neural network is trained a preset number of times to obtain the spectrum sensing neural network after preliminary training.

[0062] S4. All cognitive users perform full-band perception within the frequency band to obtain the spectrum occupancy status of any time slot. All cognitive users select channels based on the greedy criterion and communicate on their selected channels to obtain an average frequency utilization success rate.

[0063] Specifically, all cognitive users perform full-band sensing within the frequency band and obtain the spectrum occupancy status of the kth time slot. Where n = 1, ..., N, represents one of the N cognitive users. The cognitive user selects a channel based on the greedy criterion, which is the decision criterion that maximizes the result. All cognitive users select channels based on the greedy criterion. And communicate on the channels of their respective choices to obtain the average frequency success rate The frequency utilization success rate is an indicator of whether the signal power of a user after accessing a channel can meet the communication service quality. The frequency utilization success rate can be expressed as

[0064]

[0065] Where θ∈(0.9,1) is a constant. The closer the value of θ is to 1, the higher the accuracy of the averaging will be but more iterations will be required. The smaller the value, the greater the randomness but fewer iterations will be required. You can choose the corresponding value of θ according to the actual needs. Indicates whether the frequency usage of the tth time slot is successful. If the frequency usage is successful, If the frequency fails,

[0066] All users update the spectrum perception neural network based on the communication results. Users can obtain the predicted value of the spectrum occupancy status according to the spectrum perception neural network and obtain the strategy of collaborative access to spectrum holes.

[0067] S5 All cognitive users determine a loss function for training the preliminarily trained spectrum sensing neural network based on the average frequency utilization success rate, calculate the gradient of the loss function, and update the preliminarily trained spectrum sensing neural network based on the gradient, and obtain the anti-interference strategy of the cognitive user based on the updated spectrum sensing neural network after training.

[0068] Specifically, all users determine a loss function for training the preliminarily trained spectrum sensing neural network based on the average frequency utilization success rate, and update the preliminarily trained spectrum sensing neural network. Users can obtain a predicted value of the spectrum occupancy state based on the trained spectrum sensing neural network, and can obtain a strategy for collaborative access to spectrum holes. The specific process of determining the loss function, calculating the gradient, and performing training based on the gradient are conventional technical means for those skilled in the art and will not be repeated here.

[0069] The present invention provides a spectrum cognitive collaborative anti-interference method, which can update the spectrum perception neural network using the perceived spectrum data based on the intelligent perception of cognitive users, and obtain a strategy for collaborative access to spectrum holes. It does not require information interaction between users during the learning process, and can quickly learn multi-user collaborative anti-interference strategies, effectively improving the collaborative anti-interference capability under complex spectrum conditions.

[0070] Optionally, the spectrum cognitive collaborative anti-interference method provided by the present invention, the spectrum perception neural network includes a pooling layer, a convolution layer and a fully connected layer, wherein the pooling layer is used to reduce redundant information in the spectrum data, reduce the matrix dimension and thus reduce the amount of calculation, the convolution layer is used to extract the characteristics of signal occupancy in the spectrum data, and the fully connected layer is used to fit the frequency utilization strategy.

[0071] Specifically, spectrum data is a two-dimensional matrix containing time and frequency information, that is, the present invention adopts the structure of a convolutional neural network as a spectrum perception neural network. The spectrum perception neural network includes a pooling layer, a convolution layer, and a fully connected layer. The pooling layer is mainly used to reduce redundant information in the spectrum data, reduce the matrix dimension, and thus reduce the amount of calculation; the number of convolution layers is determined according to the time-frequency two-dimensional matrix dimension of the spectrum data, and is mainly used to extract the characteristics of signal occupancy in the spectrum data. If the number of convolution layers is more, the extraction accuracy is higher, but the amount of calculation will increase significantly; the fully connected layer mainly fits the frequency strategy function, and its number of layers is determined according to the dimension of the spectrum data and the complexity of the scene. Similarly, the more layers, the higher the accuracy, and the amount of calculation will also increase.

[0072] The input of the neural network is a two-dimensional spectrum data matrix obtained by sensing the previous T time slots, and the output is an M-dimensional vector corresponding to M channel actions, representing the value estimate of each channel. The larger the value, the greater the probability that the next time slot will be idle.

[0073] The spectrum cognitive collaborative anti-interference method provided by the present invention, the spectrum perception neural network can obtain the spectrum data prediction of the next time slot based on the input historical perception spectrum data by setting the pooling layer, the convolution layer and the fully connected layer.

[0074] Optionally, in the spectrum cognitive collaborative anti-interference method provided by the present invention, in step S2, all cognitive users label the perceived spectrum data and store it in a data pool, and randomly batch sample spectrum data from the data pool to train the spectrum perception neural network, specifically including:

[0075] Label the perceived spectrum data and obtain the labeled spectrum data in is the spectrum occupancy status of the t-th time slot, is an M-dimensional vector, In the tth time slot, cognitive user n perceives the occupancy state of the Mth channel; the output of the spectrum sensing neural network is Define the loss function as The gradient of the loss function is obtained, and the spectrum sensing neural network is trained using a gradient-based optimization method.

[0076] Specifically, based on the perceived spectrum data, the signal power in each channel can be calculated. If the signal power in the channel is greater than the preset power threshold, the channel is considered to be occupied. If the signal power in the channel is not greater than the preset power threshold, the channel is considered to be idle. In the tth time slot, the spectrum data can be labeled according to the occupied state or idle state of the channel to obtain the labeled spectrum data. in is an M-dimensional vector. If the mth channel is occupied, then The value is 1, if the mth channel is idle, then The value is 0.

[0077] The training of spectrum sensing neural network is completed based on the labeled spectrum data. The output of spectrum sensing neural network is Define the loss function The spectrum sensing neural network can be trained by finding the gradient of the loss function and using a gradient-based optimization method.

[0078] The spectrum cognitive collaborative anti-interference method provided by the present invention labels spectrum data and completes the training of the spectrum perception neural network based on the labeled spectrum data, so as to make the prediction of the spectrum perception neural network more accurate.

[0079] Optionally, in the spectrum-aware collaborative anti-interference method provided by the present invention, in step S4, all cognitive users select channels based on a greedy criterion and communicate on their respective selected channels to obtain an average frequency utilization success rate, specifically including:

[0080] All cognitive users select channels based on the greedy criterion, that is, input the spectrum data of the kth time slot into the spectrum sensing neural network. Spectrum sensing neural network output Select the action corresponding to the maximum value as the result of the decision, that is

[0081] Among them, the frequency success rate is in

[0082]

[0083] Where θ∈(0.9,1) is a constant, Indicates whether the frequency usage of the kth time slot is successful. If the frequency usage is successful, If the frequency fails,

[0084] Specifically, based on the trained spectrum sensing neural network, the spectrum data of the current k-th time slot can be input into it. The output of the spectrum sensing neural network can be obtained Select the action corresponding to the largest value as the decision result. You can obtain spectrum sensing prediction results.

[0085] Get the spectrum occupancy status of the kth time slot Where n = 1, ..., N, represents one of the N cognitive users. The cognitive user selects a channel based on the greedy criterion, which is the decision criterion that maximizes the result. All cognitive users select channels based on the greedy criterion. And communicate on the channels of their respective choices to obtain the average frequency success rate The frequency utilization success rate is an indicator of whether the signal power of a user after accessing a channel can meet the communication service quality. The frequency utilization success rate can be expressed as

[0086]

[0087] Where θ∈(0.9,1) is a constant. The closer the value of θ is to 1, the higher the accuracy of the averaging will be but more iterations will be required. The smaller the value, the greater the randomness but fewer iterations will be required. You can choose the corresponding value of θ according to the actual needs. Indicates whether the frequency usage of the tth time slot is successful. If the frequency usage is successful, If the frequency fails,

[0088] The spectrum cognitive collaborative anti-interference method provided by the present invention is based on the prediction results of the spectrum perception neural network and adopts a greedy criterion to select channels, so that all cognitive users can choose a better strategy and have stronger anti-interference ability.

[0089] Optionally, in the spectrum cognitive collaborative anti-interference method provided by the present invention, in step S5, all cognitive users determine a loss function for training the preliminarily trained spectrum sensing neural network based on the average frequency success rate, calculate the gradient of the loss function, and update the preliminarily trained spectrum sensing neural network based on the gradient, specifically including:

[0090] For cognitive user n, define the loss function The gradient of the loss function is calculated and the neural network is trained using a gradient-based optimization method.

[0091] Specifically, the spectrum sensing neural network not only processes the spectrum data but also selects the channel. Therefore, the neural network at this stage adopts the deep reinforcement learning training method. For the cognitive user n, the loss function is defined as The neural network can be trained by using gradient-based optimization methods based on the gradient of the loss function.

[0092] The spectrum cognitive collaborative anti-interference method provided by the present invention can train a neural network for a single cognitive user by finding the gradient of a loss function and optimizing the gradient, thereby ensuring that the signal-to-interference-noise ratio of the cognitive user meets the communication requirements and the accuracy of the neural network training results.

[0093] Optionally, the spectrum-aware collaborative anti-interference method provided by the present invention further includes:

[0094] Before step S1, a cognitive user is randomly selected as the control user;

[0095] After step S5, if the frequency utilization success rate of the controlled user is lower than the first threshold, all cognitive users return to step S2 in the preset time slot.

[0096] Specifically, before step S1, a cognitive user can be randomly selected as a control user to determine whether the spectrum access strategy can meet the communication needs and whether the surrounding frequency-using devices or malicious interference devices in the spectrum environment have changed the frequency usage strategy.

[0097] After step S5, it is determined whether the frequency usage success rate of the controlled user is lower than a preset first threshold. If the frequency usage success rate of the controlled user is not lower than the preset first threshold, it is considered that the frequency usage strategy does not need to be relearned; if the frequency usage success rate of the controlled user is lower than the preset first threshold, it is considered that the previous frequency usage strategy is invalid, and all cognitive users return to step S2 in the preset time slot and restart the learning of the spectrum status.

[0098] In one embodiment, the average frequency utilization success rate of the control user in the kth time slot is The maximum average frequency success rate in history is Then if Exceeds the preset reduction threshold or below the threshold It is considered that the surrounding frequency-using devices or malicious interference devices in the spectrum environment have changed the frequency usage strategy, and the spectrum environment needs to be re-learned. At this time, the control user broadcasts the control signal to all cognitive users through the control channel, and agrees that the left and right cognitive users will restart the spectrum status learning after a certain time slot.

[0099] The present invention provides a spectrum-aware collaborative anti-interference method that can promptly detect and re-learn the spectrum status when surrounding frequency-using devices or malicious interference devices change their frequency usage strategies. This method can effectively counteract the interference of malicious interference devices and obtain an efficient and reliable collaborative anti-interference communication strategy through independent learning.

[0100] The present invention also provides a multi-user anti-interference frequency game method based on the above-mentioned spectrum cognitive collaborative anti-interference method, comprising:

[0101] Establish a multi-user anti-interference frequency game model, which includes traditional communication equipment, surrounding frequency-using equipment, malicious interference equipment and multiple cognitive users. The frequency-using strategy of traditional communication equipment is π C , the frequency usage strategy of the surrounding frequency-using devices is π A , the frequency strategy of the malicious interference device is π J , the frequency strategy of cognitive user n is π n , cognitive user n is one of multiple cognitive users 1, 2, ..., N;

[0102] The model is G={N,S,A n , r n}, where N is the total number of cognitive users, S is the set of spectrum states obtained by all cognitive users at a certain moment, and A n is the action set of cognitive user n, r n is the reward function of cognitive user n;

[0103] The average frequency success rate of cognitive user n in the tth time slot As the reward of cognitive user n in time slot t Average frequency success rate for

[0104]

[0105] Where θ∈(0.9,1) is a constant, Indicates whether the frequency usage of the tth time slot is successful. If the frequency usage is successful, If the frequency fails,

[0106] Cognitive user n obtains frequency usage strategy based on spectrum sensing neural network;

[0107] For cognitive user n, the multi-user anti-interference frequency game model is constructed to meet the following conditions:

[0108] maxη n

[0109] st(π n ∩π c )=0

[0110] The cognitive user n selects a frequency utilization strategy that meets the above conditions based on a multi-user anti-interference frequency utilization game model.

[0111] Specifically, if Figure 2 As shown, the multi-user anti-interference frequency game method provided by the present invention is based on the following application scenario: in a distributed multi-user wireless communication network, multiple cognitive users share spectrum resources for wireless communication. Each cognitive user has the ability to independently perceive the spectrum, learn frequency usage strategies, and make frequency usage decisions. It can optimize its frequency usage strategy in real time based on the spectrum environment to improve communication success rates. Outside the wireless network, three types of frequency-using devices—traditional communication devices, surrounding frequency-using devices, and malicious interference devices—can affect the communications of wireless network users.

[0112] Establish a multi-user anti-interference frequency game model, which includes traditional communication equipment and multiple cognitive users. The frequency strategy of traditional communication equipment is π C , the frequency usage strategy of the surrounding frequency-using devices is π A , the frequency strategy of the malicious interference device is π J , multiple cognitive users communicate independently and distributedly by searching for spectrum holes. The frequency strategy of cognitive user n is known to be π n Cognitive user n is one of multiple cognitive users 1, 2, ..., N. The number of service channels is M, which are primarily used by each user to transmit service data. In addition to the service channels, the multi-user communication system uses a control channel to control the communication process. All cognitive users communicate according to synchronized time slots and continuously transmit signals.

[0113] The model is G={N,S,A n , r n}, where N is the total number of cognitive users, S is the set of spectrum states obtained by all cognitive users at a certain moment, and A n is the action set of cognitive user n, r n is the reward function of cognitive user n;

[0114] The average frequency success rate of cognitive user n in the tth time slot As the reward of cognitive user n in time slot t Average frequency success rate for

[0115]

[0116] Where θ∈(0.9,1) is a constant that can be selected according to the actual situation. The closer the value is to 1, the higher the accuracy of the average is, but more iterations are required. The smaller the value is, the greater the randomness is. Indicates whether the frequency usage of the tth time slot is successful. If the frequency usage is successful, If the frequency fails,

[0117] The cognitive user n obtains the frequency utilization strategy based on the spectrum sensing neural network.

[0118] For cognitive user n, the multi-user anti-interference frequency game model is constructed to meet the following conditions:

[0119] maxη n

[0120] st(π n ∩π c )=0

[0121] That is, the average frequency utilization success rate is maximized under the condition of pre-planning the frequency utilization strategy to avoid traditional communication equipment.

[0122] The multi-user anti-interference frequency game method provided by the present invention can establish a multi-user anti-interference frequency game model, conduct statistics on spectrum interference conditions and select the best frequency utilization strategy, maximize the average frequency utilization success rate, and effectively improve the collaborative anti-interference capability under complex spectrum conditions.

[0123] Optionally, the multi-user anti-interference frequency game method provided by the present invention is: Indicates whether the frequency usage of the tth time slot is successful, including:

[0124]

[0125] Where δ(X) is the indicator function. If the X condition is met, then δ(X) = 1. If the X condition is not met, then δ(X) = 0. is the signal-to-interference-and-noise ratio (SIR) received by cognitive user n, i.e., the ratio of the sum of the signal, interference, and noise, which is related to the frequency utilization strategies adopted by the frequency-using devices and malicious interference devices around the t-th time slot. th is the communication service quality requirement threshold.

[0126] Specifically, δ(X) is an indicator function. If the X condition is satisfied, δ(X) = 1. If the X condition is not satisfied, δ(X) = 0. In the present invention, it is to judge whether the X condition is satisfied. Greater than λ th ,like Greater than λ th ,but is 1, if No greater than λ th ,but is 0.

[0127] is the signal-to-interference-noise ratio received by cognitive user n, that is, The received useful user signal power is the sum of the power of the other three harmful interferences plus the noise power, λ th is the communication service quality requirement threshold. If the signal-to-interference-noise ratio reaches the service quality requirement, the communication is considered successful. is 1 if the value is set, otherwise it is 0.

[0128] The multi-user anti-interference frequency game method provided by the present invention determines whether the signal transmission of the cognitive user meets the communication service quality by judging the size of the signal-to-interference-plus-noise ratio received by the cognitive user and the communication service quality requirement threshold, thereby ensuring that the signal transmission of the cognitive user can meet the communication quality requirements in the multi-user anti-interference frequency game model.

[0129] The spectrum-aware collaborative anti-interference method provided by the present invention is described below through a specific embodiment.

[0130] The system simulation uses Python language, and the system parameters can be set according to the actual situation. The parameter setting does not affect the generality. The parameter settings are: the number of users is, the user shared frequency band is 20MHz, the number of channels is M=20, the user power is 0dBm, the power of other frequency devices is 20dBm, and the communication service quality threshold of the cognitive user is λ th =10dB. Spectrum sensing is performed every 1ms with a resolution of 100kHz and 200 sampling points. Each spectrum data point stores 200ms of spectrum sensing results. The length of each time slot for a cognitive user is 5ms. The malicious interference mode is sweeping interference, which periodically and sequentially interferes with all channels at a sweep rate of 2GHz / s. The spectrum sensing neural network consists of one pooling layer, two convolutional layers, and two fully connected layers. The pooling layer uses the max-pooling algorithm with a 2×2 filter. The two convolutional layers contain 16 4×4 convolution kernels with a stride of 4 and 32 3×3 convolution kernels with a stride of 2, respectively. The number of neurons in the two fully connected layers is 256 and 128, respectively. Reinforcement learning discount factor. Figure 3This is a simulation graph showing how the average frequency utilization success rate of the algorithm proposed in this embodiment of the present invention changes with the number of iterations. As can be seen from the graph, the proposed algorithm can quickly converge to a value close to 1 in complex interference environments. This indicates that it can find an effective collaborative anti-interference strategy, ensuring that all users do not select the same channel, thus avoiding co-channel interference. It can also avoid interfering signals outside the network and complete communication tasks.

[0131] The present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above method. The computer-readable storage medium may include, but is not limited to, any type of disk, including a floppy disk, an optical disk, a DVD, a CD-ROM, a microdrive, a magneto-optical disk, a ROM, a RAM, an EPROM, an EEPROM, a DRAM, a VRAM, a flash memory device, a magnetic card or an optical card, a nanosystem (including a molecular memory IC), or any type of medium or device suitable for storing instructions and / or data.

[0132] It should be noted that for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

[0133] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0134] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interface, and the indirect coupling or communication connection of the device or unit can be electrical or other forms.

[0135] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0136] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0137] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: various media that can store program codes, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0138] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0139] The above is only an exemplary embodiment of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure herein, those skilled in the art will easily think of the implementation scheme of the present disclosure. This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art that are not recorded in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

[0140] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0141] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A spectrum-aware collaborative anti-interference method, which uses a trained spectrum-aware neural network to obtain a channel selection strategy to achieve anti-interference communication, characterized in that: The following steps are involved: S1 establishes an initialized spectrum sensing neural network for all cognitive users; S2: All cognitive users perform full-band sensing within the frequency band, acquiring spectrum data including spectrum occupancy status for any time slot. All cognitive users access channels and communicate based on preset frequency hopping rules. The spectrum data sensed by all cognitive users is labeled and stored in a data pool. Spectrum data is randomly batch-sampled from the data pool to train the spectrum sensing neural network. S3 repeats step S2 for a preset number of cycles to obtain a spectrum sensing neural network after preliminary training; In S4, all cognitive users perform full-band sensing within the frequency band to obtain the spectrum occupancy status of any time slot. All cognitive users select channels based on the greedy criterion and communicate on their selected channels to obtain the average frequency utilization success rate. S5. All cognitive users determine a loss function for training the preliminarily trained spectrum sensing neural network based on the average frequency utilization success rate, calculate the gradient of the loss function, and update the preliminarily trained spectrum sensing neural network based on the gradient; and obtain the anti-interference strategy of the cognitive user based on the trained and updated spectrum sensing neural network.

2. The spectrum-aware collaborative anti-interference method according to claim 1, wherein: The spectrum sensing neural network includes a pooling layer, a convolutional layer and a fully connected layer, wherein: The pooling layer is used to reduce redundant information in the spectrum data and reduce the matrix dimension. The convolution layer is used to extract the characteristics of signal occupancy in the spectrum data. The fully connected layer is used to fit the frequency utilization strategy.

3. The spectrum-aware collaborative anti-interference method according to claim 1, wherein: In step S2, all cognitive users label the spectrum data they perceive and store it in a data pool. Random batch sampling of spectrum data from the data pool is used to train the spectrum perception neural network, specifically including: Label the perceived spectrum data and obtain the labeled spectrum data ,in is the spectrum occupancy status of the t-th time slot, is an M-dimensional vector, In the tth time slot, cognitive user n perceives the occupancy state of the Mth channel; The output of the spectrum sensing neural network is , define the loss function as , ; The gradient of the loss function is calculated, and a gradient-based optimization method is used to train the spectrum sensing neural network.

4. The spectrum-aware collaborative anti-interference method according to claim 1, wherein: In step S4, all cognitive users select channels based on the greedy criterion and communicate on the channels they select to obtain an average frequency utilization success rate, which specifically includes: All cognitive users select channels based on the greedy criterion, that is, input the spectrum data of the kth time slot into the spectrum sensing neural network. , spectrum sensing neural network output , Select the action corresponding to the maximum value as the result of the decision, that is ; Among them, the frequency success rate is ,in in is a constant, , represents whether the frequency usage of the kth time slot is successful. If the frequency usage is successful, If the frequency fails, .

5. The spectrum-aware collaborative anti-interference method according to claim 4, wherein: In step S5, all cognitive users determine a loss function for training the preliminarily trained spectrum sensing neural network based on the average frequency usage success rate, calculate the gradient of the loss function, and update the preliminarily trained spectrum sensing neural network based on the gradient, specifically including: For cognitive user n, define the loss function , find the gradient of the loss function and use the gradient-based optimization method to train the neural network.

6. The spectrum-aware collaborative anti-interference method according to claim 1, wherein: Also includes: Before step S1, a cognitive user is randomly selected as the control user; After step S5, if the frequency utilization success rate of the control user is lower than the first threshold, all cognitive users return to step S2 in a preset time slot.

7. A multi-user anti-interference frequency game method based on the spectrum cognitive collaborative anti-interference method according to claim 5, characterized in that: include: A multi-user anti-interference frequency game model is established, which includes traditional communication equipment, surrounding frequency-using equipment, malicious interference equipment and multiple cognitive users. The frequency-using strategy of the traditional communication equipment is: , the frequency usage strategy of surrounding frequency-using devices is , the frequency strategy of malicious interference equipment is , the frequency usage strategy of cognitive user n is , the cognitive user n is one of a plurality of cognitive users 1, 2, ..., N; The model is , where N is the total number of cognitive users, S is the set of spectrum states obtained by all cognitive users at a certain moment, is the action set of the cognitive user n, is the reward function of cognitive user n; The average frequency success rate of cognitive user n in the tth time slot As the reward of cognitive user n in time slot t , the average frequency success rate for in is a constant, , represents whether the frequency usage of the tth time slot is successful. If the frequency usage is successful, If the frequency fails, ; The cognitive user n obtains a frequency utilization strategy based on the spectrum sensing neural network; For cognitive user n, the multi-user anti-interference frequency game model is constructed to meet the following conditions: The cognitive user n selects a frequency utilization strategy that meets the above conditions based on the multi-user anti-interference frequency utilization game model.

8. The multi-user anti-interference frequency game method according to claim 7, characterized in that: described , indicating whether the frequency usage of the tth time slot is successful, specifically including: in is the indicator function. If the X condition is met, then =1, if the X condition is not met, then =0; is the signal-to-interference-and-noise ratio received by cognitive user n, which is related to the frequency usage strategies adopted by frequency-using devices and malicious interference devices around the tth time slot. is the communication service quality requirement threshold.

9. A spectrum-aware collaborative anti-interference device, characterized in that: The method comprises at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit executes the steps of the method according to any one of claims 1 to 6.

10. A storage medium, characterized in that: It stores a computer program executable by an access authentication device. When the computer program runs on the access authentication device, the access authentication device executes the steps of the method according to any one of claims 1 to 6.

Images