Machine learning based signal encryption effectiveness identification method, medium and device

Abstract

This invention provides a method, medium, and apparatus for identifying the effectiveness of signal encryption based on machine learning. The method includes the following steps: Step 1: Obtaining bitstream data through signal preprocessing; Step 2: Performing randomness detection on the bitstream data to construct a random feature vector; Step 3: Training and validating an SVM classification model using the random feature vector, and using the validated SVM classification model to complete the identification of signal encryption effectiveness. This invention addresses the problem of signal overt transmission caused by the wireless communication equipment itself or human error. By employing machine learning-based extraction and analysis of signal overt and cover features, it effectively solves the problems of high complexity, large sample size, and high data quality requirements of traditional methods.

Description

Technical Field

[0001] This invention relates to the field of signal encryption technology, and more specifically, to a method, medium, and apparatus for identifying the effectiveness of signal encryption based on machine learning. Background Technology

[0002] Currently, cryptography is deeply integrated into various information systems and platforms. The public disclosure of important information may lead to financial losses or privacy breaches. Effective signal encryption is a crucial means of ensuring information security. To ensure wireless signal security, the most direct and effective method is to conduct rigorous encryption effectiveness testing, thereby identifying potential risks and vulnerabilities in the cryptographic information security defense system. Currently, there are three main methods for testing the effectiveness of wireless signal encryption.

[0003] The first is the cooperation method, which requires the communication parameters of the device under test during testing. Figure 1 Phase 3 determines the encryption validity of the air interface signal by analyzing the decoded bitstream data. The advantage of this method is that it can obtain complete encrypted data for analysis and is relatively easy to implement. The disadvantages are high preconditions, limited versatility, high cost, and poor timeliness. The second method is blind decoding of the wireless signal. This method requires first parsing the parameters of the received signal, then demodulating and decoding the signal, analyzing and stripping the message protocol, and finally... Figure 1 Phase 3 determines the encryption effectiveness of the air interface signal by analyzing the decoded bitstream data. Its advantages include low security requirements (no need to know the parameters of the object under test) and high versatility. Its disadvantages include significant difficulty in blind signal parsing and low timeliness. Thirdly, it directly... Figure 1 Phase 1 uses machine learning methods to directly extract the electromagnetic features of the air interface signal to determine the signal clarity and achieve encryption validity detection. The advantages of this method are low requirements for security conditions (no need to know the parameters of the object under test), strong versatility and ease of use, and high timeliness. The disadvantages are that it requires a large amount of data, has high requirements for the signal-to-noise ratio of the received signal, and the communication signal characteristics may change over time, causing inaccurate identification. Summary of the Invention

[0004] This invention aims to address the problem of signal transmission caused by wireless communication devices themselves or human error by providing a machine learning-based method for identifying the effectiveness of signal encryption. By extracting and analyzing the signal transmission characteristics based on machine learning, the effectiveness of signal encryption can be identified, effectively solving the problems of high complexity, large sample size, and high data quality requirements of traditional methods. This invention provides a signal encryption validity identification method based on machine learning, comprising the following steps: Step 1: Obtain bitstream data through signal preprocessing; Step 2: Perform randomness detection on the bitstream data and construct a random feature vector; Step 3: Train and validate the SVM classification model using random feature vectors, and use the validated SVM classification model to identify the effectiveness of signal encryption.

[0005] Furthermore, step one includes: The acquired plain signal is modulated, and after the modulation mode and parameters are identified, the acquired plain signal is demodulated according to the identified modulation mode and parameters to obtain bit stream data.

[0006] Furthermore, step two includes: Randomness is checked on the demodulated bitstream data, and the P-value obtained from the randomness detection is vectorized to construct a random feature vector.

[0007] Furthermore, step three includes: The random feature vectors obtained in step two are divided into three parts according to their functions: training data, validation data, and test data. The training and test data are imported into the SVM classification model for training and classification, and the validation data is imported into the trained SVM classification model for validation.

[0008] Furthermore, the ratio of training data, validation data, and test data is set as needed.

[0009] Furthermore, in step three, a binary classifier is constructed using a linear kernel function. Through a finite number of searches in the parameter space, the optimal combination of penalty parameters and kernel function parameters is found, thereby maximizing the performance of the SVM classification model.

[0010] The present invention also provides a computer terminal storage medium storing computer terminal executable instructions, which are used to execute the above-described signal encryption validity identification method based on machine learning.

[0011] The present invention also provides a computing device, comprising: At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the aforementioned machine learning-based signal encryption validity identification method.

[0012] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. This invention directly receives and demodulates the air interface signal of the equipment under test, without needing to know the parameters of the object under test. It has strong versatility and ease of use, high timeliness, low requirements for the signal-to-noise ratio of the received signal, and small data volume. Furthermore, since this invention uses the randomness characteristics of the received signal to distinguish between clear and dense signals, and since the randomness characteristics of clear and dense signals differ greatly and change very little over time, it solves the shortcomings of existing identification methods.

[0013] 2. Compared with existing methods, this invention avoids analyzing and stripping messages, reducing algorithm complexity, and the sample length only needs to be... bits, less than the data length required to perform a simple randomness test ( This method (using only bits) does not require traversing all signal modes and rates, achieving high accuracy with only a small sample size, thus reducing computational complexity and improving recognition accuracy. To test the performance of this method, a single time period, single mode, and single rate data set was used as the sample set to train the SVM classification model. The model was then tested with multiple time periods, multiple modes, and multiple rates, achieving an accuracy of over 97%. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of the signal encryption principle.

[0016] Figure 2 This is a flowchart of a signal encryption validity identification method based on machine learning in an embodiment of the present invention.

[0017] Figure 3 This is a schematic diagram of kernel function transformation in an SVM network model. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0019] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0020] Example Design concept: In non-cooperative mode (when the carrier frequency, modulation method, signal bandwidth, symbol rate, and other parameters of the target signal are unknown), the received test signal is modulated and identified. After identifying the modulation method and parameters, the data is demodulated. Random features are extracted from the demodulated data, and a model is generated by training a support vector machine (SVM). The SVM model is then used to identify the encryption validity of the signal.

[0021] In this embodiment, a broadband signal receiving device is used to collect wireless explicit and implicit signals. First, the explicit / implicit status of the signal transmitting device is set. Each type of signal transmission mode has three modes, and each mode has five possible transmission rates. For example... Figure 2 As shown in the figure, the signal encryption validity identification method based on machine learning proposed in this embodiment includes the following steps: Step 1: Obtain bitstream data through signal preprocessing; The acquired plain signal is modulated, and after the modulation mode and parameters are identified, the acquired plain signal is demodulated according to the identified modulation mode and parameters to obtain bit stream data.

[0022] Step 2: Perform randomness detection on the bitstream data and construct a random feature vector; like Figure 2 As shown, randomness is tested on the demodulated bitstream data. The P-value (a metric for measuring the randomness of a sample) obtained from the randomness test is vectorized to construct a randomness feature vector, so as to reflect the difference in the clarity of the signal through randomness.

[0023] Step 3: Training and Validation; The random feature vectors obtained in step two are divided into three parts according to their functions: training data, validation data, and test data. The training data and test data are imported into the SVM classification model for training and classification recognition, and the validation data is imported into the trained SVM classification model for validation.

[0024] SVM, or Support Vector Machine, is a machine learning method based on the principle of structural risk minimization, statistical learning theory, and VC dimension theory. It provides a unified solution for classifying finite samples. For nonlinear problems, the SVM first maps the input low-dimensional feature vectors to a high-dimensional feature space using a kernel function, and then performs nonlinear classification of the sample data in the high-dimensional space.

[0025] In the n-dimensional nonlinear original vector space, SVM is achieved through nonlinear transformation. vector x Mapping to a high-dimensional feature space, SVM solves nonlinear classification problems by constructing an optimal classification plane in that high-dimensional space. SVM needs to address two issues: [the following is unclear and likely refers to a specific problem or task]. Minimize and utilize kernel function transformations for efficient computation, such as... Figure 3 As shown.

[0026] Suppose that mapping the original low-dimensional space samples to a high-dimensional feature space H involves a nonlinear mapping. SVM will use dot product Construct a classification plane in a high-dimensional space H. If there exists a function K such that... Then it can be ignored The specific form can be achieved by simply performing a dot product operation in the high-dimensional feature space H. Based on functional theory, when the kernel function... If the Mercer condition is satisfied, then a dot product corresponding to the kernel function exists in the transformation space. Therefore, if an appropriate dot product function can be used... By constructing the optimal classification plane, linear classification can be achieved in the high-dimensional feature space through this nonlinear transformation.

[0027] This method constructs a binary classifier using a linear kernel function. Through a finite number of searches in the parameter space, it finds the optimal combination of penalty and kernel function parameters, thus maximizing the performance of the SVM classification model. Simulation results demonstrate that, combined with a random feature extraction method, SVM achieves excellent results in both accuracy and stability for signal density recognition.

[0028] Furthermore, in some embodiments, a computer terminal storage medium is proposed, storing computer terminal executable instructions for executing the machine learning-based signal encryption validity identification method described in the preceding embodiments. Examples of computer storage media include magnetic storage media (e.g., floppy disks, hard disks, etc.), optical recording media (e.g., CD-ROMs, DVDs, etc.), or memory such as memory cards, ROMs, or RAMs. The computer storage medium can also be distributed across a network-connected computer system, for example, as an application store.

[0029] Furthermore, in some embodiments, a computing device is proposed, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the machine learning-based signal encryption validity identification method as described in the foregoing embodiments. Examples of computing devices include PCs, tablet computers, smartphones, or PDAs, etc.

[0030] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A signal encryption validity identification method based on machine learning, characterized in that, Includes the following steps: Step 1: In non-cooperative mode, the modulation mode of the acquired signal under test is identified. After identifying the modulation mode and parameters, the acquired signal under test is demodulated according to the identified modulation mode and parameters to obtain bit stream data. Step 2: Perform randomness detection on the bitstream data and construct a random feature vector; Step three: train and verify the SVM classification model using the random feature vector, and use the verified SVM classification model to complete the signal encryption effectiveness identification; wherein the sample data used for training and identification has a length of 10 5 bits.

2. The signal encryption validity identification method based on machine learning according to claim 1, characterized in that, Step two includes: Randomness is checked on the demodulated bitstream data, and the P-value obtained from the randomness detection is vectorized to construct a random feature vector.

3. The signal encryption validity identification method based on machine learning according to claim 1, characterized in that, Step three includes: The random feature vectors obtained in step two are divided into three parts according to their functions: training data, validation data, and test data. The training and test data are imported into the SVM classification model for training and classification, and the validation data is imported into the trained SVM classification model for validation.

4. The signal encryption validity identification method based on machine learning according to claim 3, characterized in that, The ratio of training data, validation data, and test data is set according to requirements.

5. The signal encryption validity identification method based on machine learning according to claim 1, characterized in that, In step three, a binary classifier is constructed using a linear kernel function. Through a finite number of searches in the parameter space, the optimal combination of penalty parameters and kernel function parameters is found to optimize the performance of the SVM classification model.

6. A computer terminal storage medium storing computer terminal executable instructions, characterized in that, The computer terminal can execute instructions for performing the machine learning-based signal encryption validity identification method as described in any one of claims 1-5.

7. A computing device, characterized in that, include: At least one processor; The at least one processor is also connected in communication with a memory, wherein the memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the machine learning-based signal encryption validity identification method as described in any one of claims 1-5.

Images