Wireless information source identification method and system under industrial Internet of Things multipath effect
By generating adversarial networks and deep learning algorithms to process wireless signal characteristics, the multi-path effect is eliminated, and the accuracy of source recognition in industrial Internet of Things systems is improved, the impact of multi-path effect on source recognition is solved, and information security guarantee is enhanced .
Patent Information
- Application Number
- CN202411236457.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-04
- Publication Date
- 2024-12-24
AI Technical Summary
In industrial Internet of Things systems, the multi-path effect causes the identification and analysis of wireless signals to become complicated, affecting information security, and it is difficult for the prior art to effectively apply source identification technology.
By obtaining the characteristic information of wireless communication signals, using the generative adversarial network to eliminate interference from the multipath effect, extracting the identification gene, and inputting it into the signal classification model to identify the source information, and constructing characteristic data with deep learning algorithms and filtering and noise reduction processing Set and gene data sets, perform high-dimensional clustering and wireless channel equalization processing.
It effectively improves the accuracy of source identification of wireless communication signals, enhances the security of industrial Internet of Things information, and improves the communication monitoring capabilities of network access equipment.
Smart Images
Figure CN119179944A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the technical field of wireless signal source identification, and in particular to a method and system for wireless signal source identification under multipath effects of industrial Internet of Things. Background Art
[0002] With the rapid development of industrial Internet of Things technology, all walks of life are increasingly relying on wireless networks for communication and data exchange, and the industrial production environment is becoming increasingly digital. However, this digital trend also brings challenges in information security, especially the many potential security vulnerabilities and risks in the industrial Internet of Things system. Therefore, the research and application of information security technology has become crucial.
[0003] As one of the key technologies for security protection, source identification is widely used in industrial IoT systems to ensure the legitimacy and integrity of data sources and prevent malicious attacks and data tampering. However, the application of source identification in wireless networks still faces many challenges, one of the most important of which is the impact of multipath effects. Multipath effects cause signals to go through multiple paths during transmission, resulting in problems such as Doppler effect and delay spread, making signal identification and analysis more complex and difficult.
[0004] Therefore, how to solve information security problems in the context of the rapid development of the Industrial Internet of Things, especially how to effectively apply source identification technology when facing the challenges brought by the multipath effect, has become one of the focuses and difficulties of current research. Summary of the invention
[0005] The present disclosure aims to solve at least one of the problems existing in the prior art and provide a method and system for wireless signal source identification under multipath effect of industrial Internet of Things.
[0006] In one aspect of the present disclosure, a method for identifying a wireless signal source under a multipath effect of an industrial Internet of Things is provided, the method comprising:
[0007] Obtain wireless communication signals of the source to be identified in the industrial Internet;
[0008] Extracting characteristic information of the wireless communication signal;
[0009] Inputting the characteristic information of the wireless communication signal into a trained generator to obtain an identification gene corresponding to the wireless communication signal; the trained generator is used to eliminate the interference effect of multipath effect on the wireless communication signal;
[0010] The marker gene corresponding to the wireless communication signal is input into a trained signal classification model to obtain the information source corresponding to the wireless communication signal.
[0011] Optionally, the training process of the generator and the signal classification model includes:
[0012] Obtain the preamble signal data set corresponding to the wireless devices in the industrial Internet;
[0013] Performing filtering, noise reduction and spectrum analysis on the preamble signal data set to construct a feature data set;
[0014] Performing high-dimensional clustering on the feature data set to evaluate expected values corresponding to the feature data set under a multipath effect;
[0015] Based on the feature data set and the expected value corresponding thereto, a generative adversarial network is used to perform wireless channel equalization processing to eliminate the interference influence of multipath effect, thereby obtaining the trained generator;
[0016] Using the trained generator to extract the identification genes corresponding to the wireless devices in the industrial Internet, and construct a gene data set;
[0017] The signal classification model constructed based on a supervised deep learning algorithm is trained using the gene data set to obtain the trained signal classification model.
[0018] Optionally, the performing wireless channel equalization processing by using a generative adversarial network based on the feature data set and the corresponding expected value to eliminate the influence of multipath interference to obtain the trained generator includes:
[0019] The neural network structures and parameters of the generator and the discriminator are respectively set to construct the generative adversarial network;
[0020] Inputting the feature data set into the generator to obtain a generated sample, and inputting the generated sample and the expected value into the discriminator to obtain a corresponding recognition result;
[0021] The accuracy of the recognition result is used as the loss function of the discriminator, and the misjudgment rate of the recognition result is used as the loss function of the generator. The generative adversarial network is alternately trained to reduce the difference between the generated sample and the corresponding expected value, so as to eliminate the interference of the multipath effect and obtain the trained generator.
[0022] Optionally, the using the trained generator to extract the identification genes corresponding to the wireless devices in the industrial Internet to construct a gene data set includes:
[0023] Obtain signal characteristic information corresponding to each of the wireless devices in the industrial Internet;
[0024] Inputting each of the signal feature information into the trained generator to obtain the identification gene corresponding to each of the wireless devices;
[0025] The identification genes of all the wireless devices are merged to construct the gene data set.
[0026] Optionally, the obtaining of a preamble signal data set corresponding to a wireless device in the industrial Internet includes:
[0027] Using a signal receiver to collect multiple communication signals sent by the wireless device;
[0028] Extracting the preamble code fragments in each of the communication signals according to a preset data frame structure;
[0029] The extracted preamble code segments are merged to construct the preamble code signal data set.
[0030] Optionally, performing filtering, noise reduction and spectrum analysis on the preamble signal data set to construct a feature data set includes:
[0031] Designing a digital filter, and determining the order and cutoff frequency of the digital filter according to the noise intensity of the wireless channel; wherein the digital filter is a low-pass Butterworth digital filter;
[0032] Using the digital filter to perform noise reduction processing on the preamble signal data set;
[0033] Performing IQ demodulation on each signal data in the preamble signal data set after noise reduction processing, and extracting characteristic information of each signal data;
[0034] The characteristic information of each of the signal data is merged to construct the characteristic data set.
[0035] Optionally, performing high-dimensional clustering on the feature data set to evaluate expected values corresponding to the feature data set under a multipath effect includes:
[0036] Constructing a high-dimensional feature space on the feature data set, and clustering the feature data in the feature data set using a k-means algorithm;
[0037] Taking the minimum sum of squares within the cluster as the optimization goal, the cluster center is determined through iterative search;
[0038] The characteristic data of the cluster center point is used as the expected value corresponding to the characteristic data set under the multipath effect.
[0039] Another aspect of the present disclosure provides a wireless signal source identification system under multipath effect of industrial Internet of Things, the identification system comprising:
[0040] A signal acquisition module is used to obtain wireless communication signals of the source to be identified in the industrial Internet;
[0041] A data processing module, used to extract characteristic information of the wireless communication signal;
[0042] A gene extraction module, used to input the characteristic information of the wireless communication signal into a trained generator to obtain an identification gene corresponding to the wireless communication signal; the trained generator is used to eliminate the interference effect of multipath effect on the wireless communication signal;
[0043] The signal identification classification module is used to input the identification gene corresponding to the wireless communication signal into the trained signal classification model to obtain the source information corresponding to the wireless communication signal.
[0044] Optionally, the signal acquisition module is further used to obtain a preamble signal data set corresponding to a wireless device in the industrial Internet;
[0045] The data processing module is further used to perform filtering, noise reduction and spectrum analysis on the preamble signal data set to construct a feature data set;
[0046] The gene extraction module is also used to perform high-dimensional clustering on the feature data set to evaluate the expected value corresponding to the feature data set under the multipath effect; based on the feature data set and the expected value corresponding to it, a generative adversarial network is used to perform wireless channel equalization processing to eliminate the interference of the multipath effect, and obtain the trained generator; the trained generator is used to extract the identification genes corresponding to each wireless device in the industrial Internet, and construct a gene data set;
[0047] The signal identification classification module is also used to use the gene data set to train the signal classification model constructed based on the supervised deep learning algorithm to obtain the trained signal classification model.
[0048] Optionally, the gene extraction module is further used to perform wireless channel equalization processing using a generative adversarial network based on the feature data set and the corresponding expected value to eliminate the interference effect of multipath effects and obtain the trained generator, including:
[0049] The gene extraction module is specifically used for:
[0050] The neural network structures and parameters of the generator and the discriminator are respectively set to construct the generative adversarial network;
[0051] Inputting the feature data set into the generator to obtain a generated sample, and inputting the generated sample and the expected value into the discriminator to obtain a corresponding recognition result;
[0052] The accuracy of the recognition result is used as the loss function of the discriminator, and the misjudgment rate of the recognition result is used as the loss function of the generator. The generative adversarial network is alternately trained to reduce the difference between the generated sample and the corresponding expected value, so as to eliminate the interference of the multipath effect and obtain the trained generator.
[0053] Optionally, the gene extraction module is further used to extract the identification genes corresponding to each wireless device in the industrial Internet by using the trained generator to construct a gene data set, including:
[0054] The gene extraction module is specifically used for:
[0055] Obtain signal characteristic information corresponding to each of the wireless devices in the industrial Internet;
[0056] Inputting each of the signal feature information into the trained generator to obtain the identification gene corresponding to each of the wireless devices;
[0057] The identification genes of all the wireless devices are merged to construct the gene data set.
[0058] Optionally, the signal acquisition module is further used to obtain a preamble signal data set corresponding to a wireless device in the industrial Internet, including:
[0059] The signal acquisition module is specifically used for:
[0060] Using a signal receiver to collect multiple communication signals sent by the wireless device;
[0061] Extracting the preamble code fragments in each of the communication signals according to a preset data frame structure;
[0062] The extracted preamble code segments are merged to construct the preamble code signal data set.
[0063] Optionally, the data processing module is further used to perform filtering, noise reduction and spectrum analysis on the preamble signal data set to construct a feature data set, including:
[0064] The data processing module is specifically used for:
[0065] Designing a digital filter, and determining the order and cutoff frequency of the digital filter according to the noise intensity of the wireless channel; wherein the digital filter is a low-pass Butterworth digital filter;
[0066] Using the digital filter to perform noise reduction processing on the preamble signal data set;
[0067] Performing IQ demodulation on each signal data in the preamble signal data set after noise reduction processing, and extracting characteristic information of each signal data;
[0068] The characteristic information of each of the signal data is merged to construct the characteristic data set.
[0069] Optionally, the gene extraction module is further used to perform high-dimensional clustering on the feature data set to evaluate the expected value corresponding to the feature data set under the multi-path effect, including:
[0070] The gene extraction module is specifically used for:
[0071] Constructing a high-dimensional feature space on the feature data set, and clustering the feature data in the feature data set using a k-means algorithm;
[0072] Taking the minimum sum of squares within the cluster as the optimization goal, the cluster center is determined through iterative search;
[0073] The characteristic data of the cluster center point is used as the expected value corresponding to the characteristic data set under the multipath effect.
[0074] Another aspect of the present disclosure provides an electronic device, including:
[0075] at least one processor; and,
[0076] a memory communicatively connected to at least one processor; wherein,
[0077] The memory stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor so that the at least one processor can execute the wireless signal source identification method under the multi-path effect of the industrial Internet of Things recorded above.
[0078] Another aspect of the present disclosure provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the wireless signal source identification method under the multipath effect of the industrial Internet of Things as described above.
[0079] Compared with the prior art, the present invention can eliminate the adverse effects of the multipath effect of the wireless channel on the source identification process in the application scenario of the industrial Internet of Things, effectively improve the accuracy of source identification of wireless communication signals, thereby effectively improving the communication monitoring capability of the industrial Internet of Things network-connected equipment and strengthening the information security protection of the industrial Internet of Things. BRIEF DESCRIPTION OF THE DRAWINGS
[0080] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplifications do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.
[0081] Figure 1 A flowchart of a method for identifying wireless signal sources under multipath effects in an industrial Internet of Things provided by an embodiment of the present disclosure;
[0082] Figure 2 A training flow chart of a generator and a signal classification model provided in another embodiment of the present disclosure;
[0083] Figure 3 A schematic diagram of a wireless device communication signal and feature data set processing process provided by another embodiment of the present disclosure;
[0084] Figure 4 A schematic diagram of a generator training process and a signal classification model training process provided by another embodiment of the present disclosure;
[0085] Figure 5 A schematic diagram of the structure of a wireless signal source identification system under multipath effect of industrial Internet of Things provided by another embodiment of the present disclosure;
[0086] Figure 6 A schematic structural diagram of an electronic device provided in another embodiment of the present disclosure. DETAILED DESCRIPTION
[0087] In order to make the purpose, technical scheme and advantages of the embodiments of the present disclosure clearer, the embodiments of the present disclosure will be described in detail below in conjunction with the accompanying drawings. However, it can be understood by those skilled in the art that in each embodiment of the present disclosure, many technical details are proposed in order to enable readers to better understand the present disclosure. However, even without these technical details and various changes and modifications based on the following embodiments, the technical scheme claimed for protection in the present disclosure can also be implemented. The division of the following embodiments is for the convenience of description and should not constitute any limitation on the specific implementation of the present disclosure. The various embodiments can be combined and referenced with each other without contradiction.
[0088] One embodiment of the present disclosure relates to a method for identifying wireless signal sources under multipath effects of industrial Internet of Things, the process of which is as follows: Figure 1 As shown, including:
[0089] Step S110, obtaining a wireless communication signal of a source to be identified in the industrial Internet. For example, step S110 can collect a wireless communication signal sent by a wireless device in the industrial Internet through various signal receivers, and use the collected wireless communication signal as the wireless communication signal of the source to be identified.
[0090] Step S120, extracting characteristic information of the wireless communication signal. For example, step S120 may first filter and de-noise the wireless communication signal to reduce or eliminate the impact of noise on the wireless communication signal. After the de-noising process, step S120 may extract the amplitude, phase and other characteristics of the wireless communication signal through spectrum analysis to obtain characteristic information of the wireless communication signal.
[0091] Step S130, input the characteristic information of the wireless communication signal into the trained generator to obtain the identification gene corresponding to the wireless communication signal. The trained generator is used to eliminate the interference effect of the multipath effect on the wireless communication signal. Specifically, the trained generator can perform function mapping on the characteristic information of the wireless communication signal, thereby eliminating the interference effect of the multipath effect on the wireless communication signal, and obtain the identification gene corresponding to the wireless communication signal, and the identification gene is used to indicate the wireless device corresponding to the wireless communication signal. The trained generator can be trained in the training process of the signal classification model.
[0092] Step S140, input the identification gene corresponding to the wireless communication signal into the trained signal classification model to obtain the source information corresponding to the wireless communication signal. Specifically, the trained signal classification model can identify which wireless device in the industrial Internet the wireless communication signal comes from through the identification gene corresponding to the wireless communication signal, thereby obtaining the source information corresponding to the wireless communication signal, which is used to indicate the wireless device that sent the wireless communication signal.
[0093] Compared with the prior art, the wireless signal source identification method under the multipath effect of the industrial Internet of Things provided by the embodiment of the present disclosure can eliminate the adverse effects of the multipath effect of the wireless channel on the signal source identification process in the application scenario of the industrial Internet of Things, and effectively improve the signal source identification accuracy of the wireless communication signal, thereby effectively improving the communication monitoring capability of the industrial Internet of Things network-connected equipment and strengthening the information security protection of the industrial Internet of Things.
[0094] For example, Figure 2 As shown in Figure 1, the training process of the signal classification model includes:
[0095] Step S210, obtaining a preamble signal data set corresponding to wireless devices in the industrial Internet.
[0096] Specifically, the preamble is a special sequence sent from the sender to the receiver before data transmission starts in a wireless network for time synchronization, channel estimation and other functions. It has a fixed time period. Therefore, step S210 can construct a preamble signal data set based on the preamble information in multiple communication signals sent by any wireless device in the industrial Internet to help improve the accuracy and reliability of the signal classification model.
[0097] Exemplarily, step S210 includes: using a signal receiver to collect multiple communication signals sent by a wireless device; extracting the preamble code segments in each communication signal according to a preset data frame structure; merging the extracted preamble code segments to construct a preamble code signal data set.
[0098] Specifically, the preset data frame structure can be selected and set according to actual needs. For example, the preset data frame structure can be a data frame structure defined by the IEEE 802.11 protocol. The IEEE 802.11 protocol is a standard for wireless network communication defined by the Institute of Electrical and Electronics Engineers (IEEE) and is a common standard for wireless networks today.
[0099] For example, combining Figure 3 In step S210, any wireless device can be extracted from a number of wireless devices in the industrial Internet as a sample source, and n communication signals emitted by the extracted wireless device can be collected by using a signal receiver. The n communication signals can be expressed as s (1) 、s (2) ,...,s (n) , assuming that the number of data points of the preamble code segment in each communication signal is p, the preamble code signal data set S constructed according to each preamble code segment can be expressed as in, Represent the communication signal s (i) The 1st, 2nd, ..., pth data points of the inner preamble code segment, variable i = 1, 2, ..., n.
[0100] Step S220: filtering, denoising and spectrum analysis are performed on the preamble signal data set to construct a feature data set.
[0101] By filtering and denoising the preamble signal data set, the impact of noise can be effectively reduced or eliminated, the quality of the feature data set can be improved, and the accuracy and reliability of the signal classification model can be further improved.
[0102] Exemplarily, step S220 includes: designing a digital filter, determining the order and cutoff frequency of the digital filter according to the noise intensity of the wireless channel; wherein the digital filter is a low-pass Butterworth digital filter; performing noise reduction processing on the preamble signal data set using the digital filter; performing IQ demodulation on each signal data in the preamble signal data set after noise reduction processing, and extracting characteristic information of each signal data; merging the characteristic information of each signal data to construct a characteristic data set.
[0103] Specifically, by using a low-pass Butterworth digital filter to perform noise reduction processing on the preamble signal data set, the noise reduction effect can be effectively improved, the influence of noise can be further effectively reduced or eliminated, and the quality of the feature data set can be improved.
[0104] IQ demodulation refers to the process of separating the in-phase component (I) and the quadrature component (Q) from a complex signal, and is often used to process signals of modulation modes such as quadrature amplitude modulation (QAM) and orthogonal frequency division multiplexing (OFDM). In addition, when performing IQ demodulation on a signal, the amplitude, phase and other characteristics of the signal can also be extracted. Therefore, step S220 can extract the characteristic information of each signal data, that is, the data point in each preamble code segment, through spectrum analysis of IQ demodulation, and construct a characteristic data set based on the characteristic information to further help improve the accuracy and reliability of the signal classification model.
[0105] For example, combining Figure 3 , assuming that m features are extracted from each signal data, then the communication signal s (i) The corresponding feature information f (i) It can be expressed as The feature dataset F can be expressed as Among them, f (1) f when i=1 (i) ,Right now By analogy, f (n) f when i=n (i) ,Right now
[0106]
[0107] Step S230 , performing high-dimensional clustering on the feature data set, and evaluating the expected value corresponding to the feature data set under the multi-path effect.
[0108] Exemplarily, step S230 includes: constructing a high-dimensional feature space on the feature data set, clustering the feature data in the feature data set using the k-means algorithm; determining the cluster center through iterative search with the minimum sum of squares within the cluster as the optimization goal; and using the feature data of the cluster center as the expected value corresponding to the feature data set under the multi-path effect.
[0109] For example, combining Figure 3 , for the feature dataset For example, through high-dimensional clustering of features, the expected value corresponding to the feature data set F is obtained It can be expressed as in, Each element in represents The expected value corresponding to each element in is the characteristic expected value, where i = 1, 2, …, n.
[0110] Step S240, based on the feature data set and its corresponding expected value, a generative adversarial network is used to perform wireless channel equalization processing to eliminate the interference of multipath effects and obtain a trained generator.
[0111] Exemplarily, step S240 includes: setting the neural network structure and parameters of the generator and the discriminator respectively to construct a generative adversarial network; inputting the feature data set into the generator to obtain generated samples, and inputting the generated samples and expected values into the discriminator to obtain corresponding recognition results; using the accuracy of the recognition results as the loss function of the discriminator, and using the misjudgment rate of the recognition results as the loss function of the generator, alternately training the generative adversarial network, reducing the difference between the generated samples and their corresponding expected values, so as to eliminate the interference of the multipath effect and obtain a trained generator.
[0112] For example, combining Figure 4 , the feature dataset Input the generator, and get the n generated samples generated by the generator based on the feature data set F. The data dimension of each generated sample is consistent with the expected value. The data dimension is the same as that of Each of them includes m rows and p columns, with a total of m×p elements. The expected value corresponding to it The input discriminator performs sample recognition. Through iterative optimization of the generator's loss function and the discriminator's loss function, the generator and the discriminator are trained alternately until the discriminator cannot distinguish between the generated sample and its corresponding expected value. At this time, the generative adversarial network remaps the feature data in the feature data set, eliminating the signal feature differences caused by the multipath effect. Thus, a trained generator is obtained. The trained generator can be used in the specific application of source identification to eliminate the interference of the multipath effect on the wireless communication signal.
[0113] Step S250, using the trained generator to extract the identification gene corresponding to each wireless device in the industrial Internet, and construct a gene data set.
[0114] Exemplarily, step S250 includes: obtaining signal feature information corresponding to each wireless device in the industrial Internet; inputting each signal feature information into a trained generator to obtain an identification gene corresponding to each wireless device; merging the identification genes of all wireless devices to construct a gene data set.
[0115] Specifically, the trained generator can eliminate the adverse interference of multipath effects on communication signals. Therefore, by using the trained generator to remap the signal feature information corresponding to each wireless device, the identification gene corresponding to each wireless device when sending a communication signal can be extracted.
[0116] For example, combining Figure 4 , assuming that N device features, i.e., the signal feature information corresponding to N wireless devices, are F1, F2, ..., F N , then F1, F2, ..., F N By inputting the trained generator, we can get the corresponding identification gene of each wireless device. Marker gene By composing a data set, a gene data set can be obtained.
[0117] Step S260, using the gene data set to train the signal classification model constructed based on the supervised deep learning algorithm to obtain a trained signal classification model.
[0118] Specifically, step S260 can use each marker gene in the gene data set and a data label indicating the corresponding source type, training the signal classification model built based on the supervised deep learning algorithm to obtain a trained signal classification model, thereby applying the trained signal classification model to source identification of wireless networks in multipath environments.
[0119] In order to enable those skilled in the art to better understand the training process of the signal classification model in the above-mentioned embodiment, Figure 4 Take this as an example for specific explanation.
[0120] Step [1], from the N wireless devices in the industrial Internet, select any one wireless device as a sample source, use a signal receiver to collect n communication signals sent by the extracted wireless device, extract the preamble code fragments in each communication signal, and construct a preamble code signal data set. Then, filter, reduce noise and perform spectrum analysis on the preamble code signal data set to obtain the feature information f corresponding to the n communication signals. (1) 、f (2) , ..., f (n) , based on the feature information f (1) 、f (2) , ..., f (n) , build a feature dataset.
[0121] Step [2]: The feature information f in the feature dataset is (1) 、f (2) , ..., f (n) Input the generator with the set neural network structure and parameters to get the corresponding n generated samples
[0122] Step [3]: generate n samples and its corresponding expected value Input the discriminator with set neural network structure and parameters for sample recognition, where the expected value It is obtained by performing high-dimensional clustering on the feature dataset and evaluating it under multi-path effects.
[0123] Step [4], through iterative optimization of the generator loss function and the discriminator loss function, alternately train the generator and the discriminator until the discriminator cannot distinguish between the generated sample and its corresponding expected value. The difference between.
[0124] Step [5]: When the discriminator cannot distinguish between the generated sample and its corresponding expected value When the difference between , we can get the trained generator, that is, the trained generator.
[0125] Step [6], obtaining the signal characteristic information F1, F2, ..., F corresponding to the N wireless devices in the industrial Internet N , F1, F2, ..., F N Input the trained generator to obtain the corresponding identification gene of each wireless device Marker gene Composed data sets, and obtained gene data sets.
[0126] Step [7], using the The signal classification model constructed based on the supervised deep learning algorithm is trained to obtain a trained signal classification model, and the trained signal classification model can realize the signal source identification of the wireless network in a multipath environment.
[0127] Another embodiment of the present disclosure relates to a wireless signal source identification system under multipath effect of industrial Internet of Things, such as Figure 5 As shown, it includes a signal acquisition module 510, a data processing module 520, a gene extraction module 530, and a signal identification classification module 540.
[0128] The signal acquisition module 510 is used to obtain the wireless communication signal of the signal source to be identified in the industrial Internet.
[0129] The data processing module 520 is used to extract characteristic information of the wireless communication signal.
[0130] The gene extraction module 530 is used to input the characteristic information of the wireless communication signal into the trained generator to obtain the identification gene corresponding to the wireless communication signal; the trained generator is used to eliminate the interference effect of the multipath effect on the wireless communication signal.
[0131] The signal identification classification module 540 is used to input the identification gene corresponding to the wireless communication signal into the trained signal classification model to obtain the signal source information corresponding to the wireless communication signal.
[0132] Exemplarily, the signal acquisition module 510 is also used to obtain a preamble signal data set corresponding to wireless devices in the industrial Internet.
[0133] The data processing module 520 is also used to perform filtering, noise reduction and spectrum analysis on the preamble signal data set to construct a feature data set.
[0134] The gene extraction module 530 is also used to perform high-dimensional clustering on the feature data set and evaluate the expected values corresponding to the feature data set under the multipath effect; based on the feature data set and its corresponding expected values, a generative adversarial network is used to perform wireless channel equalization processing to eliminate the interference of the multipath effect; and a generative adversarial network is used to extract the identifier genes corresponding to each wireless device in the industrial Internet and construct a gene data set.
[0135] The signal identification classification module 540 is also used to train a signal classification model constructed based on a supervised deep learning algorithm using a gene data set to obtain a trained signal classification model.
[0136] Exemplarily, the gene extraction module 530 is further used to perform wireless channel equalization processing using a generative adversarial network based on the feature data set and its corresponding expected value to eliminate the interference effect of multipath effects, including:
[0137] The gene extraction module 530 is specifically used to: respectively set the neural network structure and parameters of the generator and the discriminator to construct a generative adversarial network; input the feature data set into the generator to obtain a generated sample, input the generated sample and the expected value into the discriminator to obtain the corresponding recognition result; use the accuracy of the recognition result as the loss function of the discriminator, and use the misjudgment rate of the recognition result as the loss function of the generator, alternately train the generative adversarial network, reduce the difference between the generated sample and its corresponding expected value, so as to eliminate the interference of the multipath effect, and obtain a trained generator.
[0138] Exemplarily, the gene extraction module 530 is also used to extract the identification genes corresponding to each wireless device in the industrial Internet by using a generative adversarial network to construct a gene data set, including:
[0139] The gene extraction module 530 is specifically used to: obtain the signal feature information corresponding to each wireless device in the industrial Internet; input each signal feature information into a trained generator to obtain the identification gene corresponding to each wireless device; merge the identification genes of all wireless devices to construct a gene data set.
[0140] Exemplarily, the signal acquisition module 510 is also used to obtain a preamble signal data set corresponding to a wireless device in the industrial Internet, including:
[0141] The signal acquisition module 510 is specifically used to: use a signal receiver to collect multiple communication signals sent by wireless devices; extract the preamble code segments in each communication signal according to a preset data frame structure; merge the extracted preamble code segments to construct a preamble code signal data set.
[0142] Exemplarily, the data processing module 520 is further used to perform filtering, noise reduction and spectrum analysis on the preamble signal data set to construct a feature data set, including:
[0143] The data processing module 520 is specifically used for: designing a digital filter, determining the order and cutoff frequency of the digital filter according to the noise intensity of the wireless channel; wherein the digital filter is a low-pass Butterworth digital filter; performing noise reduction processing on the preamble signal data set using the digital filter; performing IQ demodulation on each signal data in the preamble signal data set after noise reduction processing, and extracting characteristic information of each signal data; merging the characteristic information of each signal data to construct a characteristic data set.
[0144] Exemplarily, the gene extraction module 530 is further used to perform high-dimensional clustering on the feature data set and evaluate the expected value corresponding to the feature data set under the multi-path effect, including:
[0145] The gene extraction module 530 is specifically used for: constructing a high-dimensional feature space on the feature data set, clustering the feature data in the feature data set using the k-means algorithm; taking the minimum sum of squares within the cluster as the optimization goal, determining the cluster center through iterative search; and using the feature data of the cluster center as the expected value corresponding to the feature data set under the multi-path effect.
[0146] The specific implementation method of the wireless signal source identification system under the multi-path effect of the industrial Internet of Things provided in the embodiment of the present disclosure can be found in the wireless signal source identification method under the multi-path effect of the industrial Internet of Things provided in the embodiment of the present disclosure, and will not be repeated here.
[0147] Compared with the prior art, the wireless signal source identification system under the multipath effect of the industrial Internet of Things provided by the embodiments of the present disclosure can eliminate the adverse effects of the multipath effect of the wireless channel on the signal source identification process in the application scenario of the industrial Internet of Things, and effectively improve the signal source identification accuracy of the wireless communication signal, thereby effectively improving the communication monitoring capability of the industrial Internet of Things network-connected equipment and strengthening the information security protection of the industrial Internet of Things.
[0148] Another embodiment of the present disclosure relates to an electronic device, such as Figure 6 As shown, including:
[0149] at least one processor 601; and,
[0150] A memory 602 is communicatively connected to at least one processor 601; wherein,
[0151] The memory 602 stores instructions that can be executed by at least one processor 601, and the instructions are executed by at least one processor 601 so that the at least one processor 601 can execute the wireless source identification method under the multi-path effect of the industrial Internet of Things described in the above embodiment.
[0152] Among them, the memory and the processor are connected in a bus manner, and the bus may include any number of interconnected buses and bridges, and the bus connects various circuits of one or more processors and memories together. The bus can also connect various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be one element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices on a transmission medium. The data processed by the processor is transmitted on a wireless medium via an antenna, and further, the antenna also receives data and transmits the data to the processor.
[0153] The processor is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory can be used to store data used by the processor when performing operations.
[0154] Another embodiment of the present disclosure relates to a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the method for identifying a wireless signal source under multi-path effects of the industrial Internet of Things described in the above embodiment is implemented.
[0155] That is, those skilled in the art can understand that all or part of the steps in the method described in the above embodiments can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including several instructions for making a device (which can be a single-chip microcomputer, chip, etc.) or a processor (processor) perform all or part of the steps of the method described in each embodiment of the present disclosure. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.
[0156] Those skilled in the art will appreciate that the above-mentioned embodiments are specific embodiments for implementing the present disclosure, and in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present disclosure.
Claims
1. A wireless signal source identification method under multipath effect of industrial Internet of Things, characterized in that: The identification method comprises: Obtain wireless communication signals of the source to be identified in the industrial Internet; Extracting characteristic information of the wireless communication signal; Inputting the characteristic information of the wireless communication signal into a trained generator to obtain an identification gene corresponding to the wireless communication signal; the trained generator is used to eliminate the interference effect of multipath effect on the wireless communication signal; The marker gene corresponding to the wireless communication signal is input into a trained signal classification model to obtain the information source corresponding to the wireless communication signal.
2. The identification method according to claim 1, characterized in that: The training process of the generator and the signal classification model includes: Obtain the preamble signal data set corresponding to the wireless devices in the industrial Internet; Performing filtering, noise reduction and spectrum analysis on the preamble signal data set to construct a feature data set; Performing high-dimensional clustering on the feature data set to evaluate expected values corresponding to the feature data set under a multipath effect; Based on the feature data set and the expected value corresponding thereto, a generative adversarial network is used to perform wireless channel equalization processing to eliminate the interference influence of multipath effect, thereby obtaining the trained generator; Using the trained generator to extract the identification genes corresponding to the wireless devices in the industrial Internet, and construct a gene data set; The signal classification model constructed based on a supervised deep learning algorithm is trained using the gene data set to obtain the trained signal classification model.
3. The identification method according to claim 2, characterized in that: The method of performing wireless channel equalization processing based on the feature data set and the expected value corresponding thereto by using a generative adversarial network to eliminate the influence of multipath interference and obtain the trained generator includes: The neural network structures and parameters of the generator and the discriminator are respectively set to construct the generative adversarial network; Inputting the feature data set into the generator to obtain a generated sample, and inputting the generated sample and the expected value into the discriminator to obtain a corresponding recognition result; The accuracy of the recognition result is used as the loss function of the discriminator, and the misjudgment rate of the recognition result is used as the loss function of the generator. The generative adversarial network is alternately trained to reduce the difference between the generated sample and the corresponding expected value, so as to eliminate the interference of the multipath effect and obtain the trained generator.
4. The identification method according to claim 2, characterized in that: The method of using the trained generator to extract the identification genes corresponding to the wireless devices in the industrial Internet and constructing a gene data set includes: Obtain signal characteristic information corresponding to each of the wireless devices in the industrial Internet; Inputting each of the signal feature information into the trained generator to obtain the identification gene corresponding to each of the wireless devices; The identification genes of all the wireless devices are merged to construct the gene data set.
5. The identification method according to any one of claims 2 to 4, characterized in that: The method of obtaining a preamble signal data set corresponding to a wireless device in the industrial Internet includes: Using a signal receiver to collect multiple communication signals sent by the wireless device; Extracting the preamble code fragments in each of the communication signals according to a preset data frame structure; The extracted preamble code segments are merged to construct the preamble code signal data set.
6. The identification method according to any one of claims 2 to 4, characterized in that: The filtering, denoising and spectrum analysis are performed on the preamble signal data set to construct a feature data set, including: Designing a digital filter, and determining the order and cutoff frequency of the digital filter according to the noise intensity of the wireless channel; wherein the digital filter is a low-pass Butterworth digital filter; Using the digital filter to perform noise reduction processing on the preamble signal data set; Performing IQ demodulation on each signal data in the preamble signal data set after noise reduction processing, and extracting characteristic information of each signal data; The characteristic information of each of the signal data is merged to construct the characteristic data set.
7. The identification method according to any one of claims 2 to 4, characterized in that: The performing high-dimensional clustering on the feature data set to evaluate the expected value corresponding to the feature data set under the multi-path effect includes: Constructing a high-dimensional feature space on the feature data set, and clustering the feature data in the feature data set using a k-means algorithm; Taking the minimum sum of squares within the cluster as the optimization goal, the cluster center is determined through iterative search; The characteristic data of the cluster center point is used as the expected value corresponding to the characteristic data set under the multipath effect.
8. A wireless signal source identification system under multipath effect of industrial Internet of Things, characterized in that: The identification system comprises: A signal acquisition module is used to obtain wireless communication signals of the source to be identified in the industrial Internet; A data processing module, used to extract characteristic information of the wireless communication signal; A gene extraction module, used to input the characteristic information of the wireless communication signal into a trained generator to obtain an identification gene corresponding to the wireless communication signal; the trained generator is used to eliminate the interference effect of multipath effect on the wireless communication signal; The signal identification classification module is used to input the identification gene corresponding to the wireless communication signal into the trained signal classification model to obtain the source information corresponding to the wireless communication signal.
9. An electronic device, characterized in that: include: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the wireless source identification method under multi-path effects of the industrial Internet of Things according to any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for identifying wireless information sources under multipath effects of the industrial Internet of Things according to any one of claims 1 to 7 is implemented.