Base station signal recovery method for realizing interference suppression based on deep neural network

Abstract

The invention discloses a base station signal recovery method for realizing interference suppression based on a deep neural network. The method comprises the following steps: S1, constructing a communication scene of a base station; s2, traversing all combinations of the interference signals selected by each jammer by the edge server, and constructing a sample set under all the combinations under each combination; s3, constructing a base station signal recovery model based on a deep neural network algorithm, training the base station signal recovery model by using the signal samples in the sample set to obtain a trained base station signal recovery model, and transmitting the trained base station signal recovery model to a base station B; and S4, the base station A sends a communication signal to the base station B, the edge server selects an interference signal generated by each jammer and transmits the interference signal to the base station B and each jammer, and the base station B recovers the communication signal from the received signal based on the base station recovery signal model. According to the method, the number of the jammers is increased, and the deep neural network is combined, so that the recovery of the base station signal can be realized under the condition of randomly selecting the jammer interference signal combination.

Description

Technical Field

[0001] The present invention relates to the field of base station communications, and in particular to a base station signal recovery method based on deep neural networks to achieve interference suppression. Background Art

[0002] In the field of base station communications, interference is often used to prevent eavesdropping devices from obtaining communication signals. The specific approach of this solution is: a jammer generates an interference signal known to the communication base station, and then interferes with the communication signal. When the base station receives the signal, it can recover the communication signal from the received signal because it knows the interference signal. However, the eavesdropping device cannot extract the communication signal from the received signal because it does not know the interference signal.

[0003] However, a single jammer is easy to crack. Once the eavesdropping device is able to crack the jamming signal and its channel, it is still possible to recover the communication signal from the combined signal of the communication signal and the jamming signal, thereby causing information leakage. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a base station signal recovery method based on deep neural network to achieve interference suppression. By increasing the number of jammers and combining deep neural networks, the base station signal can be recovered by randomly selecting a combination of jammer interference signals.

[0005] The object of the present invention is achieved through the following technical solution: a base station signal recovery method based on deep neural network to achieve interference suppression, comprising the following steps: S1. Build a base station communication scenario: The system includes a base station A, a base station B, an edge server, and multiple jammers; each of the jammers can generate multiple interference signals. When base station A sends a desired signal to base station B, the edge server selects the interference signal generated by each jammer and transmits it to base station B and each jammer. Each jammer generates an interference signal simultaneously and transmits it in a coordinated manner according to the selection of the edge server, thereby interfering with the desired signal sent by base station A, thereby preventing eavesdropping devices from eavesdropping on the desired signal. When base station B receives the signal, it recovers the desired signal from the received signal based on the interference signals of each jammer, completing the signal transmission between base stations A and B. S2. The edge server traverses all combinations of interference signals selected by each jammer and collects information in the base station communication scenario under each combination to construct a sample set under all combinations; S3. Build a base station signal recovery model based on a deep neural network algorithm, train the base station signal recovery model using signal samples in the sample set, obtain the trained base station signal recovery model, and transmit it to base station B; S4. In an actual communication scenario, base station A sends a communication signal to base station B. At the same time, the edge server selects the interference signals generated by each jammer and transmits them to base station B and each jammer. Base station B recovers the communication signal from the received signal based on the base station signal recovery model.

[0006] The beneficial effects of the present invention are: (1) by using multiple jammers to transmit jamming signals simultaneously, the present invention can increase the complexity of the jamming signal and prevent information leakage caused by cracking by eavesdropping equipment; (2) The present invention combines a neural network approach to train a base station signal recovery model. By selecting a combination of interference signals sent by multiple jammers, the present invention obtains the signal characteristics of the received signal, known channel information, time delay information, and the selected interference signal combination, and the relationship between the signal characteristics and the communication signal through model training. Then, during actual communication, the present invention obtains the signal characteristics of the received signal, known channel information, time delay information, and the selected interference signal combination through model training, and can directly recover the base station signal. With the help of a neural network, signal recovery can be made simpler and more effective. (3) In actual communication scenarios, the present invention randomly selects a combination of interference signals by the edge server, which increases the randomness of the interference signals, further increases the difficulty of cracking the interference signals, and improves the security of communication. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0008] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0009] like Figure 1 As shown, the base station signal recovery method for implementing interference suppression based on a deep neural network includes the following steps: S1. Build a base station communication scenario: The system includes a base station A, a base station B, an edge server, and multiple jammers; each of the jammers can generate multiple interference signals. When base station A sends a desired signal to base station B, the edge server selects the interference signal generated by each jammer and transmits it to base station B and each jammer. Each jammer generates an interference signal simultaneously and transmits it in a coordinated manner according to the selection of the edge server, thereby interfering with the desired signal sent by base station A, thereby preventing eavesdropping devices from eavesdropping on the desired signal. When base station B receives the signal, it recovers the desired signal from the received signal based on the interference signals of each jammer, completing the signal transmission between base stations A and B. S2. The edge server traverses all combinations of interference signals selected by each jammer and collects information in the base station communication scenario under each combination to construct a sample set under all combinations; S3. Build a base station signal recovery model based on a deep neural network algorithm, train the base station signal recovery model using signal samples in the sample set, obtain the trained base station signal recovery model, and transmit it to base station B; S4. In an actual communication scenario, base station A sends a communication signal to base station B. At the same time, the edge server selects the interference signals generated by each jammer and transmits them to base station B and each jammer. Base station B recovers the communication signal from the received signal based on the base station signal recovery model.

[0010] In an embodiment of the present application, each jammer includes a carrier frequency and a plurality of different baseband jamming signals; The edge server selects the baseband interference signal of each jammer, so that each jammer generates a corresponding interference signal for transmission.

[0011] The step S2 comprises: S201. Assume that there are M jammers, each of which includes N types of jammer signals. The N types of jammer signals of the i-th jammer are respectively recorded as: ;in, represents the jth interference signal of the i-th jammer, i=1,2,…,M;j=1,2,…,N; Therefore, when M jammers transmit jamming signals simultaneously, there are M*N combinations of jamming signals. S202. When M jammers use any combination of interference signals, information is collected in the base station communication scenario to construct a sample set under the current interference signal combination: The step S202 includes: A1. Base station A generates a test signal , and transmit it to base station B. At the same time, M jammers generate corresponding interference signals according to the combination selected by the edge server. At this time, the signal received by base station B is recorded as , base station B transmits the received signal to the edge server; In the edge server, the interference signals of the selected M jammers are recorded as ,in represents the jamming signal of the i-th jammer selected in the current combination, , i=1,2,…,M; Assume that the channels from M jammers to base station B are respectively denoted as , the transmission delay is ,in, represents the channel from the i-th jammer to base station B, represents the propagation delay from the i-th jammer to the base station B, i=1,2,…,M; At the same time, it is known that the channel from base station A to base station B is , the propagation delay from base station A to base station B is ; A2. Construct the signal sample under the current interference signal combination (L, ): The sample characteristics are recorded as: The sample labels are recorded as: ; A3. Repeat steps A1 and A2 multiple times, add them into a set to obtain a sample set under the current interference signal combination.

[0012] S203. When the M jammers use each interference signal combination, repeat steps S201 to S202 to obtain a sample set under each interference signal combination, and add the signal samples in the sample sets under all interference signal combinations into the same set to form sample sets under all combinations.

[0013] In step S3, when the base station signal recovery model is trained using the signal samples in the sample set, the sample features in the sample set are used as the model input in each training process, and the sample labels corresponding to the sample features are used as the expected output of the model for training; after the training of all signal samples in the sample set is completed, it is considered that a trained base station signal recovery model is obtained.

[0014] The step S4 comprises: S401. The edge server randomly selects jammer interference signals and records them as ,in represents the interference signal selected by the i-th jammer in the current combination, , i=1,2,…,M; transmit the selected interference signal to base station B and the corresponding jammer respectively; S402. Base station A generates a communication signal and transmits it. Each jammer sends an interference signal from the server. The signal received by base station B is recorded as ; S403. Base station B builds signal characteristics : ; S404. Base station B will construct signal characteristics The trained base station signal recovery model is input, and the base station signal recovery model recovers the communication signal.

[0015] The foregoing description shows and describes a preferred embodiment of the present invention. However, as previously stated, it should be understood that the present invention is not limited to the form disclosed herein and should not be construed as excluding other embodiments. Instead, the present invention is applicable to various other combinations, modifications, and environments and is capable of modification within the scope of the inventive concept described herein, through the teachings above, or through techniques or knowledge in the relevant art. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention are intended to be within the scope of the appended claims.

Claims

1. A base station signal recovery method for interference suppression based on a deep neural network, characterized by: The following steps are involved: S1. Build a base station communication scenario: The system includes a base station A, a base station B, an edge server, and multiple jammers; each of the jammers can generate multiple interference signals. When base station A sends a desired signal to base station B, the edge server selects the interference signal generated by each jammer and transmits it to base station B and each jammer. Each jammer generates an interference signal simultaneously and transmits it in a coordinated manner according to the selection of the edge server, thereby interfering with the desired signal sent by base station A, thereby preventing eavesdropping devices from eavesdropping on the desired signal. When base station B receives the signal, it recovers the desired signal from the received signal based on the interference signals of each jammer, completing the signal transmission between base stations A and B. S2. The edge server traverses all combinations of interference signals selected by each jammer and collects information in the base station communication scenario under each combination to construct a sample set under all combinations; S3. Build a base station signal recovery model based on a deep neural network algorithm, train the base station signal recovery model using signal samples in the sample set, obtain the trained base station signal recovery model, and transmit it to base station B; S4. In an actual communication scenario, base station A sends a communication signal to base station B. At the same time, the edge server selects the interference signals generated by each jammer and transmits them to base station B and each jammer. Base station B recovers the communication signal from the received signal based on the base station signal recovery model.

2. The base station signal recovery method for interference suppression based on a deep neural network according to claim 1, characterized in that: Each jammer consists of a carrier frequency and multiple different baseband jamming signals; The edge server selects the baseband interference signal of each jammer, so that each jammer generates a corresponding interference signal for transmission.

3. The base station signal recovery method for interference suppression based on a deep neural network according to claim 1, characterized in that: The step S2 comprises: S201. Assume that there are M jammers, each of which includes N types of jammer signals. The N types of jammer signals of the i-th jammer are respectively recorded as: ;in, represents the jth interference signal of the i-th jammer, i=1,2,…,M;j=1,2,…,N; Therefore, when M jammers transmit jamming signals simultaneously, there are M*N combinations of jamming signals. S202. When M jammers use any combination of interference signals, information is collected in the base station communication scenario to construct a sample set under the current interference signal combination: S203. When the M jammers use each interference signal combination, repeat steps S201 to S202 to obtain a sample set under each interference signal combination, and add the signal samples in the sample sets under all interference signal combinations into the same set to form sample sets under all combinations.

4. The base station signal recovery method for interference suppression based on a deep neural network according to claim 3, characterized in that: The step S202 includes: A1. Base station A generates a test signal , and transmit it to base station B. At the same time, M jammers generate corresponding interference signals according to the combination selected by the edge server. At this time, the signal received by base station B is recorded as , base station B transmits the received signal to the edge server; In the edge server, the interference signals of the selected M jammers are recorded as ,in represents the jamming signal of the i-th jammer selected in the current combination, , i=1,2,…,M; Assume that the channels from M jammers to base station B are respectively denoted as , the transmission delay is ,in, represents the channel from the i-th jammer to base station B, represents the propagation delay from the i-th jammer to the base station B, i=1,2,…,M; At the same time, it is known that the channel from base station A to base station B is , the propagation delay from base station A to base station B is ; A2. Construct the signal sample under the current interference signal combination (L, ): The sample characteristics are recorded as: The sample labels are recorded as: ; A3. Repeat steps A1 and A2 multiple times, add them into a set to obtain a sample set under the current interference signal combination.

5. The base station signal recovery method for interference suppression based on a deep neural network according to claim 4, characterized in that: In step S3, when the base station signal recovery model is trained using the signal samples in the sample set, the sample features in the sample set are used as the model input in each training process, and the sample labels corresponding to the sample features are used as the expected output of the model for training; After all signal samples in the sample set are trained, it is considered that a trained base station signal recovery model is obtained.

6. The base station signal recovery method for interference suppression based on a deep neural network according to claim 1, characterized in that: The step S4 comprises: S401. The edge server randomly selects jammer interference signals and records them as ,in represents the interference signal selected by the i-th jammer in the current combination, , i=1,2,…,M; transmit the selected interference signal to base station B and the corresponding jammer respectively; S402. Base station A generates a communication signal and transmits it. Each jammer sends an interference signal from the server. The signal received by base station B is recorded as ; S403. Base station B builds signal characteristics : ； S404. Base station B will construct signal characteristics The trained base station signal recovery model is input, and the base station signal recovery model recovers the communication signal.

Images